AI Email Editor

AI Email Editor — independent reviews, comparisons, pricing and step-by-step guides on Aizhi.

• 

  Fillrate

  In computer graphics, a video card's pixel fillrate refers to the number of pixels that can be rendered on the screen and written to video memory in one second. Pixel fillrates are given in megapixels per second or in gigapixels per second (in the case of newer cards), and are obtained by multiplying the number of render output units (ROPs) by the clock frequency of the graphics processing unit (GPU) of a video card. A similar concept, texture fillrate, refers to the number of texture map elements (texels) the GPU can map to pixels in one second. Texture fillrate is obtained by multiplying the number of texture mapping units (TMUs) by the clock frequency of the GPU. Texture fillrates are given in mega or gigatexels per second. However, there is no full agreement on how to calculate and report fillrates. Another possible method is to multiply the number of pixel pipelines by the GPU's clock frequency. The results of these multiplications correspond to a theoretical number. The actual fillrate depends on many other factors. In the past, the fillrate has been used as an indicator of performance by video card manufacturers such as ATI and NVIDIA, however, the importance of the fillrate as a measurement of performance has declined as the bottleneck in graphics applications has shifted. For example, today, the number and speed of unified shader processing units has gained attention. Although fillrate doesn't provide a substantial bottleneck in games, it can still provide a bottleneck for certain parts of the game, for example applying a gaussian blur can be bottlenecked by fillrate. Scene complexity can be increased by overdrawing, which happens when an object is drawn to the frame buffer, and another object (such as a wall) is then drawn on top of it, covering it up. The time spent drawing the first object is thus wasted because it is not visible. When a sequence of scenes is extremely complex (many pixels have to be drawn for each scene), the frame rate for the sequence may drop. When designing graphics intensive applications, one can determine whether the application is fillrate-limited (or shader limited) by seeing if the frame rate increases dramatically when the application runs at a lower resolution or in a smaller window. Although this is not a full-proof method, modern videogame engines can dynamically reduce the level-of-detail required and thereby reducing fillrate-limited applications. The best way to find fillrate bottlenecks is to use GPU vendor software like NVIDIA Nsight Graphics, AMD Radeon GPU Profile and the Intel Graphics Performance Analyzers.

  Read more →
• 

  International Aerial Robotics Competition

  The International Aerial Robotics Competition (IARC) is a university-based robotics competition held on the campus of the Georgia Institute of Technology, currently hosted by RoboNation. Since 1991, collegiate teams with the backing of industry and government have fielded autonomous flying robots in an attempt to perform missions requiring robotic behaviors not previously exhibited by a flying machine. The term “aerial robotics” was coined by competition creator Robert Michelson in 1990 to describe a new class of small highly intelligent flying machines. Successive years of competition saw these aerial robots grow from vehicles that could barely maintain themselves in the air, to automatons which are self-stable, self-navigating, and able to interact with their environment. The goal of the competition has been to provide a reason for the state-of-the-art of aerial robotics to move forward. Challenges have been geared towards producing advances. From 1991 through 2009, six missions were proposed. Each involved fully autonomous robotic behavior undemonstrated at the time. In October 2013 a seventh mission was proposed. It was the first to involve interaction between aerial robots and multiple ground robots. In 2016, the competition and its creator were recognized during the Georgia legislative session in the form of a senate resolution as the longest running aerial robotics competition in the world. == History == === First mission === The initial mission to move a metallic disc from one side of an arena to the other was seen by many as almost impossible. The college teams improved their entries over the next two years when the competition saw its first autonomous takeoff, flight, and landing by a team from the Georgia Institute of Technology. In 1995, a team from Stanford University was able to acquire a single disk and move it from one side of the arena to the other in a fully autonomous flight—half. === Second mission === The competition mission was toughened and made less abstract by requiring teams to search for a toxic waste dump, map the location of partially buried randomly oriented toxic waste drums, identify the contents of each drum from the hazard labels on the outside of each drum, and bring a sample back from one of the drums. In 1996, a team from the Massachusetts Institute of Technology and Boston University, with backing from Draper Labs, created a small fully autonomous flying robot that repeatedly and correctly mapped the location of all five of the toxic waste drums, and correctly identified the contents of two from the air, completing approximately seventy five percent of the mission. The following year, an aerial robot developed by a team from Carnegie Mellon University completed the entire mission. === Third mission === The third mission began in 1998. It was a search and rescue mission requiring fully autonomous robots to take off, fly to a disaster area and search amid fires, broken water mains, clouds of toxic gas, and rubble. The scenario was recreated at the U.S. Department of Energy's Hazardous Material Management and Emergency Response (HAMMER) training facility. Because of the realism of the scenario, animatrons were used instead of human actors to simulate survivors incapable of extracting themselves from the disaster area. An aerial robot from Germany's Technische Universität Berlin was able to detect and avoid all of the obstacles, identify all the dead on the ground and the survivors (distinguishing between the two based on movement), and relay pictures of the survivors along with their locations back to first responders who would attempt a rescue. This mission was completed in 2000. === Fourth mission === The fourth mission was initiated in 2001. It involved three scenarios requiring the same autonomous behavior: a hostage rescue mission where a submarine 3 kilometers off the coast must send an aerial robot to find a coastal city, identify the embassy where hostages are being held, locate valid openings in the embassy building, enter (or send in a sensor probe/subvehicle) and relay pictures of the hostages 3 km to the submarine prior to mounting an amphibious assault on the embassy to free the hostages; the discovery of an ancient mausoleum where a virus had killed the archaeological team, who had radioed that an important and undocumented tapestry was hanging inside, with 15 minutes to send an autonomous aerial robot to find the mausoleum, enter it (or send in a sensor probe/subvehicle) and relay pictures of the tapestry back prior to the destruction of the mausoleum and its contents; and an explosion at a nuclear reactor facility where scientists must send in an aerial robot to find the operating reactor building, enter the building (or send in a sensor probe/subvehicle) and relay pictures of the control panels to determine if a melt-down is imminent. All three missions involved the same elements of ingress, locating, identification, entry, and relaying pictures within 15 minutes. It was conducted at the U.S. Army's Fort Benning Soldier Battle Lab using the McKenna MOUT (Military Operations on Urban Terrain) site. The fourth mission was completed in 2008 with 27 teams who had demonstrated each of the required aerial robotic behaviors, except being able to demonstrate these behaviors in under 15 minutes—a feat considered by the judges to be inevitable given more time, and therefore no longer a significant challenge. Thus the fourth mission was terminated, $80,000 in awards distributed, and the fifth mission established. === Fifth mission === The fifth mission picked up where the fourth mission left off by demonstrating the fully autonomous aerial robotic behaviors necessary to rapidly negotiate the confined internal spaces of a structure once it has been penetrated by an air vehicle. The nuclear reactor complex explosion scenario of the fourth mission was used as the backdrop for the fifth mission. The fifth mission required a fully autonomous aerial vehicle to penetrate the structure and negotiate the more complex interior space containing hallways, small rooms, obstacles, and dead ends in order to search for a designated target without the aid of global-positioning navigational aids, and relay pictures back to a monitoring station some distance from the structure. The First Symposium on Indoor Flight Issues was held in conjunction with this 2009 IARC event. === Sixth mission === The sixth mission began in 2010 as an extension of the fifth mission theme of autonomous indoor flight behavior, however it demanded more advanced behaviors than were possible by any aerial robot extant in 2010. This espionage mission involved covertly stealing a flash drive from a particular room in a building and depositing an identical drive to avoid detection of the theft. The 2010 Symposium on Indoor Flight Issues was held concurrently at the University of Puerto Rico - Mayagüez during the 20th anniversary competition. === Seventh mission === The seventh mission began in 2014 demanding more advanced behaviors than were possible by any aerial robot extant in 2014. A single autonomous aerial robot had to herd up to 10 autonomous ground robot targets across one designated end of a 20m x 20m (65.62 feet x 65.62 feet) arena in under 10 minutes. The arena had neither walls for SLAM mapping nor GPS availability. Techniques such as optical flow or optical odometry were possible solutions to navigation within the arena. Collisions with obstacle ground robots ended the run with no score. The autonomous aerial robots interacted with the ground robots in the following way: if an aerial robot touched the ground robot on top, the ground robot would turn clockwise 45°. If the aerial robot blocked its forward motion by landing in front of it, the ground robot would reverse direction. Ground robots that feely escaped the arena, counted against the aerial robot's overall score, so the autonomous aerial robots had to decide which ground robots were in imminent danger of crossing any boundary except the designated one, and redirect them toward the designated boundary.Zhejiang University was the overall winner of Mission 7, of 52 teams from 12 nations entered as competitors. === Eighth mission === In 2018, the 8th mission was announced. Mission 8 focused on non-electronic human-machine interaction for the first time, with four aerial robots assisting humans to complete tasks that one person could not independently accomplish. The gist of mission 8 involved a swarm of autonomous aerial robots working with a human to achieve a task in the presence of hostile "Sentry aerial robots" which were trying to impede the human. In 2018, the inaugural year of mission 8, the American Venue was held on the campus of the Georgia Institute of Technology in Atlanta, Georgia, and the Asia/Pacific Venue was conducted at Beihang University in Beijing China. The following year, Mission 8 was successfully completed in Kunming China at the Yunnan Innovation

  Read more →
• 

  HYPO CBR

  HYPO is a computer program, an expert system, that models reasoning with cases and hypotheticals in the legal domain. It is the first of its kind and the most sophisticated of the case-based legal reasoners, which was designed by Kevin Ashley for his Ph.D dissertation in 1987 at the University of Massachusetts Amherst under the supervision of Edwina Rissland. HYPO's design represents a hybrid generalization/comparative evaluation method appropriate for a domain with a weak analytical theory and applies to tasks that rarely involve just one right answer. The domain covers US trade secret law, and is substantially a common law domain. Since Anglo-American common law operates under the doctrine of precedent, the definitive way of interpreting problems is of necessity and case-based. Thus, HYPO did not involve the analysis of a statute, as required by the Prolog program. Rissland and Ashley (1987) envisioned HYPO as employing the key tasks performed by lawyers when analyzing case law for precedence to generate arguments for the prosecution or the defence. HYPO was a successful example of a general category of legal expert systems (LESs), it applies artificial intelligence (A.I.) techniques to the domain of legal reasoning in patent law, implementing a case-based reasoning (CBR) system, in contrast to rule based systems like MYCIN, or mixed-paradigm systems integrating CBR with rule-based or model-based reasoning like IKBALS II. A legal case-based reasoning essentially reasons from prior tried cases, comparing the contextual information in the current input case with that of cases previously tried and entered into the system. As noted by Ashley and Rissland (1988) CBR is used to "... capture expertise in domains where rules are ill-defined, incomplete or inconsistent". The HYPO project set out to model the creation of hypotheticals in law, where no case matches well enough. HYPO uses hypotheticals for a variety of tasks necessary for good interpretation: "to redefine old situations in terms of new dimensions, to create new standard cases when an appropriate one doesn’t exist, to explore and test the limits of a concept, to refocus a case by excluding some issues and to organize or cluster cases". Hypotheticals can include facts that support two conflicting lines of reasoning. So, it makes and responds to arguments from competing viewpoints about who should win the dispute. HYPO use heuristics such as making a case weaker or stronger, making a case extreme, enabling a near-miss, disabling a near-hit to generate hypotheticals in the context of an argument by using the dimensions mechanism. Dimensions have a range of values, along which the supportive strength that may shift from one side to the other. What differentiated this expert system from others was its facility not only to return a primary to best-case response but to return near-best-fit responses also. == Components == Legal knowledge in HYPO is contained in: the case-knowledge-base (CKB) and the library of dimensions. The CKB contains HYPO's base of known cases that are highly structured objects and sub-objects both real and hypothetical in the area of trade secret law. Each case is represented as a hierarchical set of frames whose slots are important facets of the case (e.g. Plaintiff, defendant, secret knowledge, employer/employee data).Ashley’s HYPO system used a database of thirty cases in the area indexed by thirteen dimensions. A key mechanism in HYPO is a dimension i.e. a mechanism to allow retrieval from the CKB, in order to represent legal cases. Ashley's dimensions are composed of (i) prerequisites, which are a set of factual predicates that must be satisfied for the dimension to apply (ii) focal slots, which accommodate one or two of the dimension's prerequisites designated as being indicative of the case's strength along that dimension and (iii) range information, which tells how a change in focal slot value effects the strength of a party's case along a given dimension. Dimensions focus attention on important aspects of cases. In HYPO's domain of misappropriation of trade secrets the dimension called “secrets voluntary disclosed” captures the idea that the more disclosures the plaintiff has made of his/her putative secret, the less convincing is his/her argument that the defendant is responsible for letting the secret. HYPO, like any other CBR system has also the following components: Similarity/relevancy metrics: that is, standards by which to evaluate the closeness of cases, judge their relevancy to the instant case, and select “most on point” cases. Half-Order Theory of the Application Domain: that is, hierarchies and taxonomies of knowledge, especially regarding the application domain. Precedent-based argumentation abilities: that is, capabilities to generate and evaluate precedent-based arguments. Knowledge to generate hypotheticals: that is, the ability to generate hypothetical cases to deal with various circumstances, like testing the validity of an interpretation or argument by providing gedanken experiments such as test cases or to fill in a weak CKB. == Functions == HYPO's method of creating an argument and justifying a solution or position has several steps. HYPO begins its processing with the current fact situation (cfs) which is direct input by the user into HYPO's representation framework. Once the user inputs the case, HYPO begins its legal analysis. The cfc is analyzed for relevant factors. Based on these factors HYPO selects the relevant cases and produces a case-analysis-record that records which dimensions apply to the cfc and which nearly apply (i.e. are "near misses"). The combined list of applicable and near miss dimensions is called the D-list. At this point the fact gathered module may request additional information from the user in order to draw a legal conclusion. Once all the facts are in the case-positioner module it uses the case-analysis record to create the claim lattice. This is a technique that organizes the relevant retrieved cases from the point of view of the cfc and makes it easy for HYPO to ascertain the most-on point cases (mopc) and to least on-point-cases. HYPO's arguments are 3ply, leading to the construction of the skeleton of an argument: it makes a point for one side, drawing the analogy between the problem and the precedent, responds with an argument for the opponent side, endeavoring to differentiate the cited case and citing other cases as counterarguments. Then it makes a final rebuttal, attempting to differentiate the counterarguments. The claim lattice also enables the HYPO-generator module to produce legally hypotheticals. With its use of dimension-based heuristics, the HYPO-generator does a heuristic search of the space of all possible cases. Lastly, the Explanation module expands upon the argument skeleton and provides explanation and justification for the different lines of analysis and cases found by HYPO. == An intelligent legal tutoring system == Legal expert systems are specifically designed to teach an area of law and are useful for pedagogical purposes. Ashley's work was mainly concerned to build tools to help students understand legal reasoning. Explanation and argument are the bases of the case method used in many professional schools in the U.S., first introduced by the Dean of the Harvard Law School, Christopher Columbus Langdell in 1870. The case method focuses on close readings of cases and principles; it involves students in pointed Socratic dialogue and makes strong use of hypotheticals (hypos). Thus, CATO (Aleven 1997) was a research project to device and test an intelligent, case-based tutorial program for teaching law students how to argue with cases implementing the HYPO program. Within the tutor system, Ashley and Aleven (1991) proposed to leverage an understanding of legal reasoning against the standard case-based tutoring methodology. What makes this tutoring system stand out is the additional levels of abstraction involved in its results. The system presents exercises, including the facts of a problem and a set of on-line cases and instructions to make, or respond to, a legal argument about the problem. The student/user will have a set of tools to analyze the problem and fashion an answer comparing it to other cases. Instead of simply generating precedent cases, the system works to interpret student responses, comparing them against a list of possibilities and responding to student entries, for example, by citing counterexamples, and providing feedback on a student's problem solving activities with explanations of correctness or giving further hints as to what may be wrong with evaluating a student's ability to perform legal reasoning and argument, examples and follow-up assignments by employing HYPO's model of case-based structure. == HYPO’s progeny == The quality of HYPO's results speak for themselves, in that a number of sequent legal reasoning systems are either directly based upon H

  Read more →
• 

  Oxa

  Oxa (formerly Oxbotica) is an autonomous vehicle software company, headquartered in Oxfordshire, England, and founded by Paul Newman and Ingmar Posner. == History == In 2013, Newman and Posner led the RobotCar UK project as part of Oxford University's Department of Engineering Science Mobile Robotics Group. RobotCar became the first autonomous vehicle on UK roads. In 2014, the pair used the newly developed technology to found Oxbotica. Oxbotica has raised over $18 million to date and is backed by the IP Group, Parkwalk Advisors and AXA XL. In 2018, Uber's former EMEA business head, Fraser Robinson, was appointed to the board of directors. In May 2019, Ozgur Tohumcu replaced Dr Graeme Smith as Oxbotica's CEO. Also in 2019, the company opened an office in Toronto, Canada. In January 2021, Oxbotica announced it had raised $47 million in a Series B round. In August 2021, the company achieved a safety landmark as the first company to have its autonomy safety case assessed by BSI (British Standards Institution) against the requirements of the UK Code of Practice 2019, PAS 1881:2020 and PAS 1883:2020, certifying the safety conformity of its autonomous vehicle trials and testing. The assessment was completed as part of Project Endeavour, the UK's first multi-city demonstration of autonomous vehicle services and capability. In December 2021, Gavin Jackson was named CEO. In January 2023, the company raised $140 million in a Series C round. In May 2023, the company changed its name to Oxa. Oxa raised $103 million (£77 million) in March 2026, including $50 million from the UK National Wealth Fund. Nvidia's venture capital division, NVentures, also invested in the Series D funding round, along with existing Oxa shareholders IP Group, Australian pension fund Hostplus, and BP Ventures, a division of the UK oil company. == Technology == Oxa designs software and hardware for the conversion of industrial vehicles into autonomous ones. Its full stack, end-to-end Universal Autonomy software is both vehicle and platform-agnostic, with no dependence on external infrastructure such as GPS. It can be deployed in any environment and on any terrain. In addition to underground uses, the technology is also useful in natural canyons and forests, where GPS signals are weak or non-existent, but also in "urban canyons" — cities with tall buildings that obstruct GPS signals for proper navigation. == Public deployments == The LUTZ Pathfinder pod had its first public demonstration in February 2015 in Milton Keynes. The Government-funded project was designed to ensure that autonomous vehicles would comply with the Highway Code. The pod featured autonomous control software from Oxbotica, including 19 sensors, cameras, radar and Lidar. As part of the GATEway Project in 2017, Oxbotica trialled seven autonomous shuttle buses in Greenwich, navigating a two-mile riverside path near London's O2 Arena on a route that is also used by pedestrians and cyclists. Oxbotica ran the UK's first trial of autonomous grocery deliveries that year, with British online supermarket Ocado in London, as the next step in the GATEway Project. In 2018, Oxbotica deployed its autonomous vehicle software at London's Gatwick Airport, which subsequently became the first airport in the world to trial an autonomous shuttle service. The electric-powered vehicles transported staff via airside roads between the airport's North and South terminals. An airside trial of Oxbotica's autonomous driving technology was then successfully completed at Heathrow Airport in partnership with IAG Cargo, the first airside trial of an autonomous vehicle at a UK airport. The Oxbotica-designed CargoPod ran autonomously along a cargo route around the airside perimeter for three weeks. As part of the UK Centre for Connected and Autonomous Vehicles-funded DRIVEN project, Oxbotica is developing and deploying a fleet of Ford Fusion autonomous vehicles running in both London and Oxford on public roads, and in conjunction with its consortium partners, running real-time insurance. AXA XL is partnering with Oxbotica on the development of smart insurance products using Oxbotica's autonomy technology to improve road safety. In 2018, Oxbotica announced a partnership with London private taxi firm Addison Lee to develop and deploy autonomous taxis in the city of London by 2021. A 3D street mapping exercise was conducted in London's Canary Wharf. In 2019, Oxbotica deployed a fleet of their autonomous technology within Ford Mondeo cars on public roads in Stratford, London to test their use in city environments. The £13.2 million project is in collaboration with The DRIVEN Project to develop self-driving cars. == Awards == 2019 Royal Academy of Engineering Silver Medal - Paul Newman 2017 Financial Times ArcelorMittal Boldness in Business Award Barclays Award for Innovation 2016 Frost & Sullivan Award, Technology Leadership for Autonomous Driving Software

  Read more →
• 

  Question answering

  Question answering (QA) is a computer science discipline within the fields of information retrieval and natural language processing (NLP) that is concerned with building systems that automatically answer questions that are posed by humans in a natural language. A question-answering implementation, usually a computer program, may construct its answers by querying a structured database of knowledge or information, usually a knowledge base. More commonly, question-answering systems can pull answers from an unstructured collection of natural language documents. Some examples of natural language document collections used for question answering systems include reference texts, compiled newswire reports, Wikipedia pages and other World Wide Web pages. == History == Two early question answering systems were BASEBALL and LUNAR. BASEBALL answered questions about Major League Baseball over a period of one year. LUNAR answered questions about the geological analysis of rocks returned by the Apollo Moon missions. Both question answering systems were very effective in their chosen domains. LUNAR was demonstrated at a lunar science convention in 1971 and it was able to answer 90% of the questions in its domain that were posed by people untrained on the system. Further restricted-domain question answering systems were developed in the following years. The common feature of all these systems is that they had a core database or knowledge system that was hand-written by experts of the chosen domain. The language abilities of BASEBALL and LUNAR used techniques similar to ELIZA and DOCTOR, the first chatterbot programs. SHRDLU was a successful question-answering program developed by Terry Winograd in the late 1960s and early 1970s. It simulated the operation of a robot in a toy world (the "blocks world"), and it offered the possibility of asking the robot questions about the state of the world. The strength of this system was the choice of a very specific domain and a very simple world with rules of physics that were easy to encode in a computer program. In the 1970s, knowledge bases were developed that targeted narrower domains of knowledge. The question answering systems developed to interface with these expert systems produced more repeatable and valid responses to questions within an area of knowledge. These expert systems closely resembled modern question answering systems except in their internal architecture. Expert systems rely heavily on expert-constructed and organized knowledge bases, whereas many modern question answering systems rely on statistical processing of a large, unstructured, natural language text corpus. The 1970s and 1980s saw the development of comprehensive theories in computational linguistics, which led to the development of ambitious projects in text comprehension and question answering. One example was the Unix Consultant (UC), developed by Robert Wilensky at U.C. Berkeley in the late 1980s. The system answered questions pertaining to the Unix operating system. It had a comprehensive, hand-crafted knowledge base of its domain, and it aimed at phrasing the answer to accommodate various types of users. Another project was LILOG, a text-understanding system that operated on the domain of tourism information in a German city. The systems developed in the UC and LILOG projects never went past the stage of simple demonstrations, but they helped the development of theories on computational linguistics and reasoning. Specialized natural-language question answering systems have been developed, such as EAGLi for health and life scientists. Question answering systems have been extended in recent years to encompass additional domains of knowledge For example, systems have been developed to automatically answer temporal and geospatial questions, questions of definition and terminology, biographical questions, multilingual questions, and questions about the content of audio, images, and video. Current question answering research topics include: interactivity—clarification of questions or answers answer reuse or caching semantic parsing answer presentation knowledge representation and semantic entailment social media analysis with question answering systems sentiment analysis utilization of thematic roles Image captioning for visual question answering Embodied question answering In 2011, Watson, a question answering computer system developed by IBM, competed in two exhibition matches of Jeopardy! against Brad Rutter and Ken Jennings, winning by a significant margin. Facebook Research made their DrQA system available under an open source license. This system uses Wikipedia as knowledge source. The open source framework Haystack by deepset combines open-domain question answering with generative question answering and supports the domain adaptation of the underlying language models for industry use cases. Large Language Models (LLMs)[36] like GPT-4[37], Gemini[38] are examples of successful QA systems that are enabling more sophisticated understanding and generation of text. When coupled with Multimodal[39] QA Systems, which can process and understand information from various modalities like text, images, and audio, LLMs significantly improve the capabilities of QA systems. == Types == Question-answering research attempts to develop ways of answering a wide range of question types, including fact, list, definition, how, why, hypothetical, semantically constrained, and cross-lingual questions. Answering questions related to an article in order to evaluate reading comprehension is one of the simpler form of question answering, since a given article is relatively short compared to the domains of other types of question-answering problems. An example of such a question is "What did Albert Einstein win the Nobel Prize for?" after an article about this subject is given to the system. Closed-book question answering is when a system has memorized some facts during training and can answer questions without explicitly being given a context. This is similar to humans taking closed-book exams. Closed-domain question answering deals with questions under a specific domain (for example, medicine or automotive maintenance) and can exploit domain-specific knowledge frequently formalized in ontologies. Alternatively, "closed-domain" might refer to a situation where only a limited type of questions are accepted, such as questions asking for descriptive rather than procedural information. Question answering systems in the context of machine reading applications have also been constructed in the medical domain, for instance related to Alzheimer's disease. Open-domain question answering deals with questions about nearly anything and can only rely on general ontologies and world knowledge. Systems designed for open-domain question answering usually have much more data available from which to extract the answer. An example of an open-domain question is "What did Albert Einstein win the Nobel Prize for?" while no article about this subject is given to the system. Another way to categorize question-answering systems is by the technical approach used. There are a number of different types of QA systems, including: rule-based systems, statistical systems, and hybrid systems. Rule-based systems use a set of rules to determine the correct answer to a question. Statistical systems use statistical methods to find the most likely answer to a question. Hybrid systems use a combination of rule-based and statistical methods. == Architecture == As of 2001, question-answering systems typically included a question classifier module that determined the type of question and the type of answer. Different types of question-answering systems employ different architectures. For example, modern open-domain question answering systems may use a retriever-reader architecture. The retriever is aimed at retrieving relevant documents related to a given question, while the reader is used to infer the answer from the retrieved documents. Systems such as GPT-3, T5, and BART use an end-to-end architecture in which a transformer-based architecture stores large-scale textual data in the underlying parameters. Such models can answer questions without accessing any external knowledge sources. == Methods == Question answering is dependent on a good search corpus; without documents containing the answer, there is little any question answering system can do. Larger collections generally mean better question answering performance, unless the question domain is orthogonal to the collection. Data redundancy in massive collections, such as the web, means that nuggets of information are likely to be phrased in many different ways in differing contexts and documents, leading to two benefits: If the right information appears in many forms, the question answering system needs to perform fewer complex NLP techniques to understand the text. Correct answers can be filtered from false positives because the syst

  Read more →
• 

  Murderbot (TV series)

  Murderbot is an American science fiction action comedy television series created by Paul Weitz and Chris Weitz for Apple TV+. It is based on All Systems Red, the first book of the series The Murderbot Diaries by Martha Wells, who serves as a consulting producer. The series stars Alexander Skarsgård as the titular character. The first season premiered on May 16, 2025 and received positive reviews. In July 2025, the series was renewed for a second season. == Premise == A media-obsessed private security construct (manufactured from cloned human tissue and mechanical parts) calling itself Murderbot must hide its newly acquired autonomy while completing dangerous assignments and being simultaneously drawn to humans, and appalled by their weakness. == Cast and characters == === Main === Alexander Skarsgård as Murderbot Noma Dumezweni as Ayda Mensah, a terraforming specialist, the President of Preservation Alliance and the leader of the science team protected by Murderbot David Dastmalchian as Gurathin, a tech expert and augmented human Sabrina Wu as Pin-Lee, a scientist and legal counsel to the team Akshay Khanna as Ratthi, a wormhole expert Tamara Podemski as Bharadwaj, a geochemist Tattiawna Jones as Arada, a biologist === Recurring === Cast of show-within-a-show The Rise and Fall of Sanctuary Moon John Cho as Eknie Jef Chem (playing Captain Hossein) Jack McBrayer as Breiller MocJac (playing Navigation Officer Hordööp-Sklanch) Clark Gregg as Arletty (playing Lieutenant Kullervv) DeWanda Wise as Pordron Bretney III Roche (playing NawBot 337 Alt 66) === Guest === Anna Konkle as Leebeebee, a member of another survey team on the planet. The character does not appear in the novella. Amanda Brugel as GrayCris Blue Leader David Reale as GrayCris Yellow == Episodes == == Production == The book series was optioned in the late 2010s, and its film adaptation was considered. In 2021, book series author Martha Wells said that a potential TV series adaptation was in development and that she had read the script and was "really excited about it". The series was green lit by Apple TV+ in 2022, with Wells serving as a consulting producer. The production design team, led by Sue Chan, started work in the autumn. Tommy Arnold, the Murderbot Diaries special edition illustrator, created the concept art for the show. After the casting was delayed by the 2023 SAG-AFTRA strike, in December 2023 it was announced that Alexander Skarsgård would produce and star in the series. He developed the character and the world of Murderbot with the showrunners. In February 2024, David Dastmalchian and Noma Dumezweni joined the cast. In March, Sabrina Wu, Tattiawna Jones, Akshay Khanna, and Tamara Podemski joined the cast. On July 10, 2025, the series was renewed for a second season. Showrunners Chris and Paul Weitz suggested the second season would combine the next three books of the series and will have longer episodes. === Filming === Principal photography for the first season took place from March–June 2024, in Toronto and parts of Ontario, Canada. Most of the filming was done on location, with the Sanctuary Moon scenes filmed on a virtual production stage. Principal photography for the second season began in mid-2026, in Madrid, Spain. It is planned to last 71 days, with Martha Wells also visiting the set. == Release == The first two episodes of Murderbot premiered on Apple TV+ on May 16, 2025, with subsequent episodes released weekly. The first season consists of ten episodes. == Reception == Even before the release of the show, numerous media sources had commented on the titular character as being coded as autistic and agender. On the review aggregator website Rotten Tomatoes, Murderbot has an approval rating of 96% with an average score of 7.5/10, based on 76 critics' reviews. The website's critical consensus states, "Alexander Skarsgård's superbly dry wit brings a lot of heart to Murderbot, making for a refreshingly jaunty sci-fi saga about finally coming out of one's shell". Metacritic, which uses a weighted average, assigned a score of 70 out of 100, based on 28 critics, indicating "generally favorable" reviews. Some reviewers have criticized Murderbot's changes to Wells' original books. Angela Watercutter of Wired noted that the series has significant tonal differences from the books and noted the show's changes to characters, particularly Murderbot and Dr. Mensah, and Wells' social commentary. === Accolades === Murderbot was a finalist for the 2025 Dragon Award for Best Science Fiction or Fantasy TV Series. Tommy Arnold won the 2025 Concept Art Association Award in the category of Live-Action Series Character Art for his work on Murderbot. Alexander Skarsgård was nominated for a Critics' Choice Award for Best Actor in a Comedy Series. Carrie Grace and Laura Jean Shannon were nominated for a Costume Designers Guild Award in the category of Excellence in Sci-Fi/Fantasy Television for their work on FreeCommerce. Amanda Jones was nominated for a Composers & Lyricists Award for Outstanding Original Title Sequence for a Television Production.

  Read more →
• 

  Agents of S.H.I.E.L.D. season 4

  The fourth season of the American television series Agents of S.H.I.E.L.D., based on the Marvel Comics spy organization S.H.I.E.L.D., follows Phil Coulson and other S.H.I.E.L.D. agents and allies after the signing of the Sokovia Accords. It is set in the Marvel Cinematic Universe (MCU) and acknowledges the continuity of the franchise's films. The season was produced by ABC Studios, Marvel Television, and Mutant Enemy Productions, with Jed Whedon, Maurissa Tancharoen, and Jeffrey Bell serving as showrunners. Clark Gregg reprises his role as Coulson from the film series, starring alongside the returning series regulars Ming-Na Wen, Chloe Bennet, Iain De Caestecker, Elizabeth Henstridge, and Henry Simmons. They are joined by John Hannah who was promoted from his recurring guest role in the third season. The fourth season was ordered in March 2016, with production taking place from that July until the following April. Due to its broadcast schedule, the season was split into three "pods": Ghost Rider for the first eight episodes, featuring recurring guest star Gabriel Luna as the supernatural Robbie Reyes / Ghost Rider and exploring mysticism in the MCU alongside the film Doctor Strange (2016); LMD, referring to the new Life Model Decoy program, for the next seven episodes which focus on recurring guest star Mallory Jansen as the LMD Aida; and Agents of Hydra for the final seven episodes, partly set in a "what if" virtual reality that allowed the return of former series regular Brett Dalton as Grant Ward. The season is also affected by the events of the film Captain America: Civil War (2016), and continues storylines established in the canceled series Agent Carter. The first episode premiered at a screening on September 19, 2016, with the season then airing for 22 episodes on ABC, from September 20, 2016, until May 16, 2017. The premiere debuted to 3.58 million viewers, down from previous season premieres but average for the series. Critical response to the season was positive, with many feeling that each pod was better than the last and in particular praising the visual effects and tone of Ghost Rider, the writing and acting of LMD, and the character development and political commentary explored during Agents of Hydra. The season saw series low viewership, but was still considered to have solved ABC's problem during its new Tuesday night timeslot, and the series was renewed for a fifth season in May 2017. == Episodes == == Cast and characters == == Production == === Development === Agents of S.H.I.E.L.D. was renewed for a fourth season on March 3, 2016, earlier than usual for the series. Executive producer Jed Whedon said on this, "We're thrilled to know going into the end of [season three] with certainty that we will be returning, because we can build our story accordingly." Executive producer Maurissa Tancharoen also noted that logistics for hiring directors for the season in advance would be easier, "which is a very nice privilege to have...that's a luxury". The end of the episode "What If..." features an onscreen tribute to Bill Paxton, who died in February 2017 and had portrayed John Garrett in the series' first season. The series paid additional tribute to Paxton in "All the Madame's Men" with promos during The Bakshi Report news segment showcasing John Garrett as a fallen American hero. The end of "World's End" features a similar onscreen tribute to Powers Boothe, who died in May 2017 and had portrayed Gideon Malick in the series' third season. === Writing === The season shifted to the later 10 pm timeslot, allowing it to take on a darker, more mature tone than previous seasons. According to Tancharoen, "The whole tagline for this year is 'Agents of S.H.I.E.L.D. After Dark'". The timeslot gave the series the opportunity to present an increased level of violence and partial nudity, as well as take more risks and present edgier themes. Following the third-season finale, Tancharoen stated that the fourth season would explore the guilt Daisy Johnson has over Lincoln Campbell's death. Executive producer Jeffrey Bell noted the writers tried to continue the tradition of "finding new combinations and new conflicts" between different sets of characters, given "a lot of procedurals [see] the same people doing the same thing for five years". Pairings that would be explored included Coulson and Mack, continuing from the end of season three, who have a mutual respect for one another due to their relationships with Daisy, and Leo Fitz and Holden Radcliffe, who work together. The Fitz-Simmons relationship was also explored more, examining the new challenges it presented for the two "working together, loving each other and living together". Following the third season's dealing with the themes of Captain America: Civil War (2016), such as the opposing reactions to the Inhumans, Whedon said that the question of "How do you deal with a war with powered people at that level, a government level?" was one that they wanted to answer in the fourth season. Tancharoen called the Inhumans "a permanent part of our universe now", with Whedon adding, "we have a quick-fire way of introducing people with powers. It gives us a lot of leeway in our world, and it lets us explore the metaphors of what it is like to be different. We will never close that chapter." With the Inhumans film being removed from Marvel Studios' release schedule, the series had "a little more freedom" and were "able to do a little bit more" with the species, including the potential of introducing some of the "classic" Inhumans, though the series would focus less on Inhumans than the third season which saw "a real significant Inhuman agenda story". It was not intended to be a spin-off of Agents of S.H.I.E.L.D. On the evolution of S.H.I.E.L.D. to featuring so many powered characters, Whedon said "the dynamic in the world has changed. There was one person with powers, and then by The Avengers there were maybe six total ... now they're much more prevalent, so there's reaction from the public based on that." The season is structured into three "pods" based on its airing schedule: the first eight episodes, subtitled Ghost Rider; LMD (Life Model Decoy) for the subsequent seven episodes; and a third pod for the final seven episodes called Agents of Hydra. Elements and characters cross over between the different pods, but the sections "definitely have a different feel" from one another, as Bell explained that 22 episodes "is a long time to hold a big bad or a single plot line, especially for an audience", and for the past two seasons, the series was able to have two separated halves that "allows us to introduce a big bad. And then, something happens and we rise somebody new ... Now, there's three of those." "Financial considerations" were also taken into account in creating the pods for the season, as using LMDs does not "cost as much as setting a guy's head on fire via CGI". In terms of writing the "complicated season", Whedon said the writers were "aware that our fans are our fans and have spent some time with these characters and are clever and see things coming sometimes ... Part of our job is to create not just what we are presenting on plot, but letting the audience be one step ahead of us and being one step ahead of that." He added that the writers knew that they wanted to tell a Ghost Rider story, an LMD story, and a "what if" scenario, and the hardest part was making each pod still fit together as a single season. The major connection ultimately became the Darkhold, which leads from the magic of Ghost Rider to the advanced science of LMD and then the Framework in Agents of Hydra. Ghost Rider also reappears in the final episode of the season, "World's End", as an additional connection. ==== Ghost Rider ==== While planning the fourth season, Marvel suggested that the series introduce Ghost Rider, after the character's film rights had returned to Marvel from Sony in May 2013. Loeb felt that this made the season unquestionably "the series' biggest" with the "most ambitious story yet". He added that "one of the things that we talked about is, S.H.I.E.L.D. always looked out for the weird, the unusual, the things that were and could be a problem for the public", and Marvel realized that Ghost Rider's abilities, which are more mystical than anything seen in the series to date, opened up "a quarter of the universe that we haven't really spent a lot of time exploring ... what happens if our very real, our very grounded agents who are very much a family have to take on something that is as bizarre and powerful and unique as Ghost Rider." Bell added that the producers would have been willing to give an entire season of the show to a Ghost Rider arc if the season was 13 episodes or less, but 22 episodes seemed too long to "feel like one flavor". The Robbie Reyes version of Ghost Rider was chosen over other versions of the character from the comics because of his relationship with his brother Gabe, w

  Read more →
• 

  Content-based image retrieval

  Content-based image retrieval, also known as query by image content (QBIC) and content-based visual information retrieval (CBVIR), is the application of computer vision techniques to the image retrieval problem, that is, the problem of searching for digital images in large databases (see this survey for a scientific overview of the CBIR field). Content-based image retrieval is opposed to traditional concept-based approaches (see Concept-based image indexing). "Content-based" means that the search analyzes the contents of the image rather than the metadata such as keywords, tags, or descriptions associated with the image. The term "content" in this context might refer to colors, shapes, textures, or any other information that can be derived from the image itself. CBIR is desirable because searches that rely purely on metadata are dependent on annotation quality and completeness. == Comparison with metadata searching == An image meta search requires humans to have manually annotated images by entering keywords or metadata in a large database, which can be time-consuming and may not capture the keywords desired to describe the image. The evaluation of the effectiveness of keyword image search is subjective and has not been well-defined. In the same regard, CBIR systems have similar challenges in defining success. "Keywords also limit the scope of queries to the set of predetermined criteria." and, "having been set up" are less reliable than using the content itself. == History == The term "content-based image retrieval" seems to have originated in 1992 when it was used by Japanese Electrotechnical Laboratory engineer Toshikazu Kato to describe experiments into automatic retrieval of images from a database, based on the colors and shapes present. Since then, the term has been used to describe the process of retrieving desired images from a large collection on the basis of syntactical image features. The techniques, tools, and algorithms that are used originate from fields such as statistics, pattern recognition, signal processing, and computer vision. === QBIC - Query By Image Content === The earliest commercial CBIR system was developed by IBM and was called QBIC (Query By Image Content). Recent network- and graph-based approaches have presented a simple and attractive alternative to existing methods. While the storing of multiple images as part of a single entity preceded the term BLOB (Binary Large OBject), the ability to fully search by content, rather than by description, had to await IBM's QBIC. === VisualRank === == Technical progress == The interest in CBIR has grown because of the limitations inherent in metadata-based systems, as well as the large range of possible uses for efficient image retrieval. Textual information about images can be easily searched using existing technology, but this requires humans to manually describe each image in the database. This can be impractical for very large databases or for images that are generated automatically, e.g. those from surveillance cameras. It is also possible to miss images that use different synonyms in their descriptions. Systems based on categorizing images in semantic classes like "cat" as a subclass of "animal" can avoid the miscategorization problem, but will require more effort by a user to find images that might be "cats", but are only classified as an "animal". Many standards have been developed to categorize images, but all still face scaling and miscategorization issues. Initial CBIR systems were developed to search databases based on image color, texture, and shape properties. After these systems were developed, the need for user-friendly interfaces became apparent. Therefore, efforts in the CBIR field started to include human-centered design that tried to meet the needs of the user performing the search. This typically means inclusion of: query methods that may allow descriptive semantics, queries that may involve user feedback, systems that may include machine learning, and systems that may understand user satisfaction levels. == Techniques == Many CBIR systems have been developed, but as of 2006, the problem of retrieving images on the basis of their pixel content remains largely unsolved. Different query techniques and implementations of CBIR make use of different types of user queries. === Query By Example === QBE (Query By Example) is a query technique that involves providing the CBIR system with an example image that it will then base its search upon. The underlying search algorithms may vary depending on the application, but result images should all share common elements with the provided example. Options for providing example images to the system include: A preexisting image may be supplied by the user or chosen from a random set. The user draws a rough approximation of the image they are looking for, for example with blobs of color or general shapes. This query technique removes the difficulties that can arise when trying to describe images with words. === Semantic retrieval === Semantic retrieval starts with a user making a request like "find pictures of Abraham Lincoln". This type of open-ended task is very difficult for computers to perform - Lincoln may not always be facing the camera or in the same pose. Many CBIR systems therefore generally make use of lower-level features like texture, color, and shape. These features are either used in combination with interfaces that allow easier input of the criteria or with databases that have already been trained to match features (such as faces, fingerprints, or shape matching). However, in general, image retrieval requires human feedback in order to identify higher-level concepts. === Relevance feedback (human interaction) === Combining CBIR search techniques available with the wide range of potential users and their intent can be a difficult task. An aspect of making CBIR successful relies entirely on the ability to understand the user intent. CBIR systems can make use of relevance feedback, where the user progressively refines the search results by marking images in the results as "relevant", "not relevant", or "neutral" to the search query, then repeating the search with the new information. Examples of this type of interface have been developed. === Iterative/machine learning === Machine learning and application of iterative techniques are becoming more common in CBIR. === Other query methods === Other query methods include browsing for example images, navigating customized/hierarchical categories, querying by image region (rather than the entire image), querying by multiple example images, querying by visual sketch, querying by direct specification of image features, and multimodal queries (e.g. combining touch, voice, etc.) == Content comparison using image distance measures == The most common method for comparing two images in content-based image retrieval (typically an example image and an image from the database) is using an image distance measure. An image distance measure compares the similarity of two images in various dimensions such as color, texture, shape, and others. For example, a distance of 0 signifies an exact match with the query, with respect to the dimensions that were considered. As one may intuitively gather, a value greater than 0 indicates various degrees of similarities between the images. Search results then can be sorted based on their distance to the queried image. Many measures of image distance (Similarity Models) have been developed. === Color === Computing distance measures based on color similarity is achieved by computing a color histogram for each image that identifies the proportion of pixels within an image holding specific values. Examining images based on the colors they contain is one of the most widely used techniques because it can be completed without regard to image size or orientation. However, research has also attempted to segment color proportion by region and by spatial relationship among several color regions. === Texture === Texture measures look for visual patterns in images and how they are spatially defined. Textures are represented by texels which are then placed into a number of sets, depending on how many textures are detected in the image. These sets not only define the texture, but also where in the image the texture is located. Texture is a difficult concept to represent. The identification of specific textures in an image is achieved primarily by modeling texture as a two-dimensional gray level variation. The relative brightness of pairs of pixels is computed such that degree of contrast, regularity, coarseness and directionality may be estimated. The problem is in identifying patterns of co-pixel variation and associating them with particular classes of textures such as silky, or rough. Other methods of classifying textures include: Co-occurrence matrix Laws texture energy Wavelet transform Orthogonal transforms (discrete Chebyshev moments) =

  Read more →
• 

  Scikit-learn

  scikit-learn (formerly scikits.learn and also known as sklearn) is a free and open-source machine learning library for the Python programming language. It features various classification, regression and clustering algorithms including support-vector machines, random forests, gradient boosting, k-means and DBSCAN, and is designed to interoperate with the Python numerical and scientific libraries NumPy and SciPy. Scikit-learn is a NumFOCUS fiscally sponsored project. == Overview == The scikit-learn project started as scikits.learn, a Google Summer of Code project by French data scientist David Cournapeau. The name of the project derives from its role as a "scientific toolkit for machine learning", originally developed and distributed as a third-party extension to SciPy. The original codebase was later rewritten by other developers. In 2010, contributors Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort and Vincent Michel, from the French Institute for Research in Computer Science and Automation in Saclay, France, took leadership of the project and released the first public version of the library on February 1, 2010. In November 2012, scikit-learn as well as scikit-image were described as two of the "well-maintained and popular" scikits libraries. In 2019, it was noted that scikit-learn is one of the most popular machine learning libraries on GitHub. At that time, the project had over 1,400 contributors and the documentation received 42 million visits in 2018. According to a 2022 Kaggle survey of nearly 24,000 respondents from 173 countries, scikit-learn was identified as the most widely used machine learning framework. == Features == Large catalogue of well-established machine learning algorithms and data pre-processing methods (i.e. feature engineering) Utility methods for common data-science tasks, such as splitting data into train and test sets, cross-validation and grid search Consistent way of running machine learning models (estimator.fit() and estimator.predict()), which libraries can implement Declarative way of structuring a data science process (the Pipeline), including data pre-processing and model fitting == Examples == Fitting a random forest classifier: == Implementation == scikit-learn is largely written in Python, and uses NumPy extensively for high-performance linear algebra and array operations. Furthermore, some core algorithms are written in Cython to improve performance. Support vector machines are implemented by a Cython wrapper around LIBSVM; logistic regression and linear support vector machines by a similar wrapper around LIBLINEAR. In such cases, extending these methods with Python may not be possible. scikit-learn integrates well with many other Python libraries, such as Matplotlib and plotly for plotting, NumPy for array vectorization, Pandas dataframes, SciPy, and many more. == History == scikit-learn was initially developed by David Cournapeau as a Google Summer of Code project in 2007. Later that year, Matthieu Brucher joined the project and started to use it as a part of his thesis work. In 2010, INRIA, the French Institute for Research in Computer Science and Automation, got involved and the first public release (v0.1 beta) was published in late January 2010. The project released its first stable version, 1.0.0, on September 24, 2021. The release was the result of over 2,100 merged pull requests, approximately 800 of which were dedicated to improving documentation. Development continues to focus on bug fixes, efficiency and feature expansion. The latest version, 1.8, was released on December 10, 2025. This update introduced native Array API support, enabling the library to perform GPU computations by directly using PyTorch and CuPy arrays. This version also included bug fixes, improvements and new features, such as efficiency improvements to the fit time of linear models. == Applications == Scikit-learn is widely used across industries for a variety of machine learning tasks such as classification, regression, clustering, and model selection. The following are real-world applications of the library: === Finance and Insurance === AXA uses scikit-learn to speed up the compensation process for car accidents and to detect insurance fraud. Zopa, a peer-to-peer lending platform, employs scikit-learn for credit risk modelling, fraud detection, marketing segmentation, and loan pricing. BNP Paribas Cardif uses scikit-learn to improve the dispatching of incoming mail and manage internal model risk governance through pipelines that reduce operational and overfitting risks. J.P. Morgan reports broad usage of scikit-learn across the bank for classification tasks and predictive analytics in financial decision-making. === Retail and E-Commerce === Booking.com uses scikit-learn for hotel and destination recommendation systems, fraudulent reservation detection, and workforce scheduling for customer support agents. HowAboutWe uses it to predict user engagement and preferences on a dating platform. Lovely leverages the library to understand user behaviour and detect fraudulent activity on its platform. Data Publica uses it for customer segmentation based on the success of past partnerships. Otto Group integrates scikit-learn throughout its data science stack, particularly in logistics optimization and product recommendations. === Media, Marketing, and Social Platforms === Spotify applies scikit-learn in its recommendation systems. Betaworks uses the library for both recommendation systems (e.g., for Digg) and dynamic subspace clustering applied to weather forecasting data. PeerIndex used scikit-learn for missing data imputation, tweet classification, and community clustering in social media analytics. Bestofmedia Group employs it for spam detection and ad click prediction. Machinalis utilizes scikit-learn for click-through rate prediction and relational information extraction for content classification and advertising optimization. Change.org applies scikit-learn for targeted email outreach based on user behaviour. === Technology === AWeber uses scikit-learn to extract features from emails and build pipelines for managing large-scale email campaigns. Solido applies it to semiconductor design tasks such as rare-event estimation and worst-case verification using statistical learning. Evernote, Dataiku, and other tech companies employ scikit-learn in prototyping and production workflows due to its consistent API and integration with the Python ecosystem. === Academia === Télécom ParisTech integrates scikit-learn in hands-on coursework and assignments as part of its machine learning curriculum. == Awards == 2019 Inria-French Academy of Sciences-Dassault Systèmes Innovation Prize: Awarded in recognition of scikit-learn's impact as a major free software breakthrough in machine learning and its role in the digital transformation of science and industry. 2022 Open Science Award for Open Source Research Software: Awarded by the French Ministry of Higher Education and Research as part of the second National Plan for Open Science. The project was recognized in the "Community" category for its technical quality, its large international contributor network, and the quality of its documentation.

  Read more →
• 

  Squirrel AI

  Squirrel Ai Learning is an international educational technology company that specializes in intelligent adaptive learning and was one of the first companies in the world to offer large scale AI-powered adaptive education solutions. == Methodology == Squirrel Ai Learning uses artificial intelligence to tailor lesson plans to each individual student. The company's AI researchers have access to the world's largest student databases, which are used to train the AI algorithms. Squirrel Ai Learning works with teachers to identify the most fine-grained possible concepts ("knowledge points") for a course in order to precisely target learning gaps. For example, middle school mathematics is broken into over 10,000 points such as rational numbers, the properties of a triangle, and the Pythagorean theorem. Each point is linked to related items, forming a "knowledge graph". Each knowledge point is addressed by videos, examples and practice problems. A textbook might address 3,000 points; ALEKS, another adaptive learning platform, uses 1,000. Each student begins with a diagnostic test to identify where to begin their learning. The system continues to refine its graph as more students proceed. Learning is not student-directed. The system decides the order of topics. == History and milestones == Squirrel Ai Learning was founded by Derek Haoyang Li in 2014. In March, 2017, The Squirrel Ai Intelligent Adaptive Learning System (IALS) was launched. IALS utilizes artificial intelligence to customize lessons, practice and evaluations for each individual student. In 2018, Squirrel Ai Learning established a joint research lab of AI adaptive learning with the institute of Automation of the Chinese Academy of Sciences. By 2019, Squirrel Ai Learning had opened 2,000 learning centers in 200 cities and registered over a million students in Asia. In 2019, Squirrel Ai Learning opened a research lab in partnership with Carnegie Mellon University. As of 2019, Squirrel Ai Learning had raised over $180 million in funding and in 2018 it surpassed $1 billion in valuation. In 2020, Squirrel Ai Learning launched the $1 million AAAI Squirrel AI Award for Artificial Intelligence for the Benefit of Humanity in partnership with AAAI. The inaugural award was given to Regina Barzilay for her work developing machine learning models to address drug synthesis and early-stage breast cancer diagnosis. In 2020, Squirrel Ai Learning established strategic partnership with DingTalk, Alibaba Group. As of 2021, Squirrel Ai Learning had served over 60,000 public schools, in over 1200 cities in Asia. Squirrel Ai plans to start offering its services in the United States in 2026. The American arm is separate from the Chinese company to avoid regulatory hurdles. As of January 2026, it had set up an "independent technology platform" in the US. == Recognition == Squirrel Ai Learning has gained recognition both in Asia and internationally including: Squirrel Ai Learning was named one of the World's Top 30 AI application case in the 2018 Synced Machine Intelligence Awards. In June 2019, Squirrel Ai Learning was named as one of the 50 smartest companies in China by MIT technology review. Squirrel Ai Learning won the GITEX 2019 Best Education Technology Award. In 2020, Squirrel Ai Learning won the UNESCO AI Innovation Award. Squirrel Ai Learning was listed in the 2020 CB Insight's AI 100, CB Insights' annual ranking of the 100 most promising AI startups in the world. Squirrel Ai Learning won Edtech Review's Best AI in Education Company of the Year award 2020.

  Read more →
• 

  Abu Dhabi Autonomous Racing League

  The Abu Dhabi Autonomous Racing League (A2RL) is an autonomous racing league based in Abu Dhabi and organized by ASPIRE, part of the UAE government's Advanced Technology Research Council. It has three distinct categories: the "car race", the drone race, and the buggy race. The first car race was held on 27 April 2024 at the Yas Marina Circuit, marking the first major autonomous formula race outside the US since the now-folded Roborace championship. The first drone race was held on 11 and 12 April 2025. == Formats == A2RL has three distinct formats, the formula racing format (dubbed the Car Race), the quadcopter drone racing format (dubbed the Drone Race), and the off-road dune buggy racing format (dubbed the Buggy Race). === Car Race === A2RL's main event, the car race is a standard formula racing format with self-driving formula cars. The cars are made by Dallara and are modified versions of Super Formula cars with Yokohama tires. These cars had the CPUs of their AIs mounted where the driver's seat is on a non-modified chassis, as well as hydraulic actuators for AI control of the vehicle, multiple sensor systems including LIDAR and GPS, and a large LED indicator showing the status of the AI. The first car race was held on 27 April 2024. This race was marked by the cars' subpar performance: Out of four cars that qualified, only two finished the race - the other two did not. The next race was held on 15 November 2025, with 11 teams. ==== Technical specifications ==== The full list of technical specifications are as follows: Chassis: Dallara EAV24 (modified Dallara SF23) Forward suspension: Pushrod type, torsion bar spring, adjustable dampers, third element Rear suspension: Pushrod type, torsion bar, coil springs, adjustable dampers, third element Tires: Yokohama Advan Drive-by-wire system: Provided by Meccanica 42, the DBW system consists of steering and brake actuators, with a central ECU that coordinates the driving actions and reacts to any critical situation in real-time. Brakes: Brembo calipers, Brembo carbon discs, electro-hydraulically activated Engine: 4 Piston Racing K20C1 (based on Honda 2.0l; turbocharged 4-cylinder engine) Gearbox: 3MO 6-speed gearbox Sensor suite: 7x Sony IMX728 cameras, 4x ZF ProWave radar units, 3x Seyond Falcon Kinetic lidar units Main computer: Neousys RGS-8805GC ==== Races held ==== === Drone Race === Created in partnership with the Drone Champions' League, the drone race is the quadcopter drone racing aerial format of the A2RL. The first race was held on 11/12 April 2025 at the ADNEC Marina Hall. 10 teams are scheduled to take part. === Buggy Race === The buggy race will be the off-road format of the A2RL using self-driving dune buggies. No date or number of teams has been announced for the first race. === Other events === A2RL is known to host AI vs AI and Human vs AI events, in Abu Dhabi and abroad. One such event took place at the Suzuka Circuit in Japan. The Human vs AI race was precluded due to AI car "Yalla" crashing into the wall during the formation lap. == Team lists ==

  Read more →
• 

  Evolutionary acquisition of neural topologies

  Evolutionary acquisition of neural topologies (EANT/EANT2) is an evolutionary reinforcement learning method that evolves both the topology and weights of artificial neural networks. It is closely related to the works of Angeline et al. and Stanley and Miikkulainen. Like the work of Angeline et al., the method uses a type of parametric mutation that comes from evolution strategies and evolutionary programming (now using the most advanced form of the evolution strategies CMA-ES in EANT2), in which adaptive step sizes are used for optimizing the weights of the neural networks. Similar to the work of Stanley (NEAT), the method starts with minimal structures which gain complexity along the evolution path. == Contribution of EANT to neuroevolution == Despite sharing these two properties, the method has the following important features which distinguish it from previous works in neuroevolution. It introduces a genetic encoding called common genetic encoding (CGE) that handles both direct and indirect encoding of neural networks within the same theoretical framework. The encoding has important properties that makes it suitable for evolving neural networks: It is complete in that it is able to represent all types of valid phenotype networks. It is closed, i.e. every valid genotype represents a valid phenotype. (Similarly, the encoding is closed under genetic operators such as structural mutation and crossover.) These properties have been formally proven. For evolving the structure and weights of neural networks, an evolutionary process is used, where the exploration of structures is executed at a larger timescale (structural exploration), and the exploitation of existing structures is done at a smaller timescale (structural exploitation). In the structural exploration phase, new neural structures are developed by gradually adding new structures to an initially minimal network that is used as a starting point. In the structural exploitation phase, the weights of the currently available structures are optimized using an evolution strategy. == Performance == EANT has been tested on some benchmark problems such as the double-pole balancing problem, and the RoboCup keepaway benchmark. In all the tests, EANT was found to perform very well. Moreover, a newer version of EANT, called EANT2, was tested on a visual servoing task and found to outperform NEAT and the traditional iterative Gauss–Newton method. Further experiments include results on a classification problem.

  Read more →
• 

  Google Cloud Dataflow

  Google Cloud Dataflow is a fully managed service for executing Apache Beam pipelines within the Google Cloud Platform ecosystem. Dataflow provides a fully managed service for executing Apache Beam pipelines, offering features like autoscaling, dynamic work rebalancing, and a managed execution environment. Dataflow is suitable for large-scale, continuous data processing jobs, and is one of the major components of Google's big data architecture on the Google Cloud Platform. At its core, Dataflow's architecture is designed to abstract away infrastructure management, allowing developers to focus purely on the logic of their data processing tasks. When a pipeline written using the Apache Beam SDK is submitted, Dataflow translates this high-level definition into an optimized job graph. The service then provisions and manages a fleet of Google Compute Engine workers to execute this graph in a highly parallelized and fault-tolerant manner. This serverless approach, combined with intelligent autoscaling of both the number of workers (horizontal) and the resources per worker (vertical), ensures that jobs have the precise amount of computational power needed at any given time, optimizing both performance and cost. The service's deep integration with the Google Cloud ecosystem makes it a powerful tool for a variety of use cases beyond simple data movement. For real-time analytics, Dataflow can ingest unbounded streams of data from Cloud Pub/Sub, perform complex transformations, and load results into BigQuery for immediate querying. In machine learning workflows, it is commonly used to preprocess and transform massive datasets stored in Cloud Storage, preparing them for training models in Vertex AI. This versatility makes it the central processing engine for modern ETL (Extract, Transform, Load) operations, streaming analytics, and large-scale data preparation within the cloud. == History == Google Cloud Dataflow was announced in June, 2014 and released to the general public as an open beta in April, 2015. In January, 2016 Google donated the underlying SDK, the implementation of a local runner, and a set of IOs (data connectors) to access Google Cloud Platform data services to the Apache Software Foundation. The donated code formed the original basis for Apache Beam. In August 2022, there was an incident where user timers were broken for certain Dataflow streaming pipelines in multiple regions, which was later resolved. Throughout 2023 and 2024, there have been various other updates and incidents affecting Google Cloud Dataflow, as documented in the release notes and service health history. The donation of the Dataflow SDK to the Apache Software Foundation was a pivotal moment, establishing Apache Beam as a unified, open-source programming model for defining both batch and streaming data pipelines. This strategic move decoupled the pipeline definition from the execution engine. As a result, developers could write portable data processing logic that was not locked into Google's ecosystem. A Beam pipeline can be executed on various runners, including Apache Flink, Apache Spark, and, of course, the highly optimized Google Cloud Dataflow service, providing flexibility and future-proofing data processing investments. == Features == Google Cloud Dataflow supports both batch and streaming data processing pipelines. It automatically handles resource provisioning, data sharding, and scaling according to workload, reducing manual configuration needed for large-scale data operations. == Use cases == Dataflow is used for ETL (Extract, Transform, Load) data pipelines, real-time analytics, and event stream processing for companies in industries such as finance, advertising, and IoT.

  Read more →
• 

  Fuzzy architectural spatial analysis

  Fuzzy architectural spatial analysis (FASA) (also fuzzy inference system (FIS) based architectural space analysis or fuzzy spatial analysis) is a spatial analysis method of analysing the spatial formation and architectural space intensity within any architectural organization. Fuzzy architectural spatial analysis is used in architecture, interior design, urban planning and similar spatial design fields. == Overview == Fuzzy architectural spatial analysis was developed by Burcin Cem Arabacioglu (2010) from the architectural theories of space syntax and visibility graph analysis, and is applied with the help of a fuzzy system with a Mamdani inference system based on fuzzy logic within any architectural space. Fuzzy architectural spatial analysis model analyses the space by considering the perceivable architectural element by their boundary and stress characteristics and intensity properties. The method is capable of taking all sensorial factors into account during analyses in conformably with the perception process of architectural space which is a multi-sensorial act.

  Read more →
• 

  Generative engine optimization

  Generative engine optimization (GEO) is one of the names given to the practice of structuring digital content and managing online presence to improve visibility in responses generated by generative artificial intelligence (AI) systems. The practice influences the way large language models (LLMs) retrieve, summarize, and present information in response to user queries. Related terms include answer engine optimization (AEO) and artificial intelligence optimization (AIO). The concept of GEO first appeared in response to generative AI technologies being integrated into mainstream search and information retrieval systems. Tools are used to monitor how websites and brands are cited, referenced, or incorporated into responses produced by large language models. == Terminology == Several overlapping terms describe related practices, and usage varies across practitioners, vendors, and publications. No consensus definition distinguishing these terms had been established in the academic literature as of early 2026, and the terms are frequently used interchangeably in trade and practitioner contexts. Other terms for the same concept include answer engine optimization (AEO), large language model optimization (LLMO), artificial intelligence optimization (AIO), and AI SEO. In 2026, Google released documentation entitled "Optimizing your website for generative AI features on Google Search." According to this documentation, "optimizing for generative AI search is optimizing for the search experience, and thus still SEO.” This position had previously been shared at conferences, with 2026 being the first time Google released official documentation stating it. == Factors influencing generative engine optimization == By early 2026, the focus of GEO practitioners shifted from simple keyword placement to "semantic relevance", a metric driven by the integration of advertising into conversational AI. OpenAI and Google began monetizing AI search results, which is not currently considered an aspect of generative engine optimization but is adjacent.

  Read more →