AI Email Editor

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

• 

  Carrier cloud

  In cloud computing, a carrier cloud is a class of cloud that integrates wide area networks (WAN) and other attributes of communications service providers’ carrier-grade networks to enable the deployment of highly-complex applications in the cloud. In contrast, classic cloud computing focuses on the data center and does not address the network connecting data centers and cloud users. This may result in unpredictable response times and security issues when business-critical data are transferred over the Internet. == History == The advent of virtualization technology, cost-effective computing hardware, and ubiquitous Internet connectivity have enabled the first wave of cloud services starting in the early years of the 21st century. But many businesses and other organizations hesitated to move to more demanding applications, from on-premises dedicated hardware to private or public clouds. As a response, communications service providers started in the 2010/2011 time frame to develop carrier clouds that address perceived weaknesses in existing cloud services. Cited weaknesses vary but often include possible downtime, security issues, high cost of custom software and data transfer, inflexibility of some cloud apps, poor customer and nonfulfillment of service level agreements (SLAs). == Characteristics == To enable the deployment of time-sensitive and business critical applications in the cloud, the carrier cloud is designed to match or even exceed the characteristics of on-premises deployments. Therefore, the carrier cloud is characterized by some or all of the following items: Configurable, elastic network performance: Typical cloud computing solutions use the best effort of the public Internet to connect cloud users and data centers. This approach provides instant connectivity but does not offer control over network capacities, latencies, and jitter. Carrier clouds address these gaps with content delivery networks and/or dedicated virtual private networks (VPN) at OSI layers 1 (optical wavelengths), 2 (data link layer), and 3 (network layer). These VPNs can be configured to offer the desired performance parameters and exhibit the same type of elasticity for the network that regular clouds provide for servers and storage. To achieve the requested performance parameters, such as low latency, cloud applications can be (automatically) allocated to distributed data centers that are close enough to the cloud users. Automatic resource placement: For a cloud with multiple data centers, information about both the data center and the connecting network is relevant for a decision of where to place cloud images and storage volumes. For this decision, carrier clouds can obtain relevant information about the network, e.g., using the Application-Layer Traffic Optimization (ALTO) protocol. High level of security and governance: Cloud application providers are subject to general and domain specific security, privacy, and governance requirements and regulations, such as the European Data Protection Directive and the U.S. Health Insurance Portability and Accountability Act. For added security, the wide area network of the carrier cloud can provide segregated encrypted or unencrypted network links that are not accessible from the general Internet. At the data center, the carrier cloud provides e.g. virtual private servers, management processes, logs, and documentation to fulfill security and governance rules. Location control: Fundamentally, cloud users should not be concerned with the geographic location of their cloud resources. However, privacy and other regulations may mandate that certain types of data must not be sent outside a national jurisdiction or other geographical region. Open APIs: Carrier clouds provide graphical user interfaces and Web application programming interfaces that allow cloud application providers to set up, manage, and monitor both, the data center and the WAN, of their cloud services. == Architecture == Carrier clouds encompass data centers at different network tiers and wide area networks that connect multiple data centers to each other as well as to the cloud users. Links between data centers are used for failover, overflow, backup, and geographic diversity. Carrier clouds can be set up as public, private, or hybrid clouds. The carrier cloud federates these cloud entities by using a single management system to orchestrate, manage, and monitor data center and network resources as a single system.

  Read more →
• 

  Mike Vernal

  Mike Vernal (born September 7, 1980) is an American business executive who is a venture capitalist at Conviction. He was previously an investor at Sequoia Capital in Silicon Valley and was one of the top executives at Facebook between 2008 and 2016. Prior to joining Sequoia Capital, he was Vice President of Search, Local, and Developer products at Facebook. == Career == Vernal joined Facebook in 2008. From 2009 to 2013, Vernal managed the Facebook Platform team and is credited with managing the Facebook Platform transition from desktop to mobile. During his time at Facebook, he served as vice president and was considered among the “top executives” who ran the company. In 2016, after eight years at Facebook, Vernal announced his plans to leave the company. In May 2016, he joined Sequoia Capital, a venture-capital firm specializing in technology startups. He is an early investor in Rippling, Clay, Notion and Statsig. In July 2023, The Information reported that Vernal was departing Sequoia. At Conviction, he has led investments in Listen Labs, OpenEvidence and Thinking Machines Lab.

  Read more →
• 

  Learning vector quantization

  In computer science, learning vector quantization (LVQ) is a prototype-based supervised classification algorithm. LVQ is the supervised counterpart of vector quantization systems. LVQ can be understood as a special case of an artificial neural network, more precisely, it applies a winner-take-all Hebbian learning-based approach. It is a precursor to self-organizing maps (SOM) and related to neural gas and the k-nearest neighbor algorithm (k-NN). LVQ was invented by Teuvo Kohonen. == Definition == An LVQ system is represented by prototypes W = ( w ( i ) , . . . , w ( n ) ) {\displaystyle W=(w(i),...,w(n))} which are defined in the feature space of observed data. In winner-take-all training algorithms one determines, for each data point, the prototype which is closest to the input according to a given distance measure. The position of this so-called winner prototype is then adapted, i.e. the winner is moved closer if it correctly classifies the data point or moved away if it classifies the data point incorrectly. An advantage of LVQ is that it creates prototypes that are easy to interpret for experts in the respective application domain. LVQ systems can be applied to multi-class classification problems in a natural way. A key issue in LVQ is the choice of an appropriate measure of distance or similarity for training and classification. Recently, techniques have been developed which adapt a parameterized distance measure in the course of training the system, see e.g. (Schneider, Biehl, and Hammer, 2009) and references therein. LVQ can be a valuable aid in classifying text documents. == Algorithm == The algorithms are presented as in. Set up: Let the data be denoted by x i ∈ R D {\displaystyle x_{i}\in \mathbb {R} ^{D}} , and their corresponding labels by y i ∈ { 1 , 2 , … , C } {\displaystyle y_{i}\in \{1,2,\dots ,C\}} . The complete dataset is { ( x i , y i ) } i = 1 N {\displaystyle \{(x_{i},y_{i})\}_{i=1}^{N}} . The set of code vectors is w j ∈ R D {\displaystyle w_{j}\in \mathbb {R} ^{D}} . The learning rate at iteration step t {\displaystyle t} is denoted by α t {\displaystyle \alpha _{t}} . The hyperparameters w {\displaystyle w} and ϵ {\displaystyle \epsilon } are used by LVQ2 and LVQ3. The original paper suggests ϵ ∈ [ 0.1 , 0.5 ] {\displaystyle \epsilon \in [0.1,0.5]} and w ∈ [ 0.2 , 0.3 ] {\displaystyle w\in [0.2,0.3]} . === LVQ1 === Initialize several code vectors per label. Iterate until convergence criteria is reached. Sample a datum x i {\displaystyle x_{i}} , and find out the code vector w j {\displaystyle w_{j}} , such that x i {\displaystyle x_{i}} falls within the Voronoi cell of w j {\displaystyle w_{j}} . If its label y i {\displaystyle y_{i}} is the same as that of w j {\displaystyle w_{j}} , then w j ← w j + α t ( x i − w j ) {\displaystyle w_{j}\leftarrow w_{j}+\alpha _{t}(x_{i}-w_{j})} , otherwise, w j ← w j − α t ( x i − w j ) {\displaystyle w_{j}\leftarrow w_{j}-\alpha _{t}(x_{i}-w_{j})} . === LVQ2 === LVQ2 is the same as LVQ3, but with this sentence removed: "If w j {\displaystyle w_{j}} and w k {\displaystyle w_{k}} and x i {\displaystyle x_{i}} have the same class, then w j ← w j − α t ( x i − w j ) {\displaystyle w_{j}\leftarrow w_{j}-\alpha _{t}(x_{i}-w_{j})} and w k ← w k + α t ( x i − w k ) {\displaystyle w_{k}\leftarrow w_{k}+\alpha _{t}(x_{i}-w_{k})} .". If w j {\displaystyle w_{j}} and w k {\displaystyle w_{k}} and x i {\displaystyle x_{i}} have the same class, then nothing happens. === LVQ3 === Initialize several code vectors per label. Iterate until convergence criteria is reached. Sample a datum x i {\displaystyle x_{i}} , and find out two code vectors w j , w k {\displaystyle w_{j},w_{k}} closest to it. Let d j := ‖ x i − w j ‖ , d k := ‖ x i − w k ‖ {\displaystyle d_{j}:=\|x_{i}-w_{j}\|,d_{k}:=\|x_{i}-w_{k}\|} . If min ( d j d k , d k d j ) > s {\displaystyle \min \left({\frac {d_{j}}{d_{k}}},{\frac {d_{k}}{d_{j}}}\right)>s} , where s = 1 − w 1 + w {\displaystyle s={\frac {1-w}{1+w}}} , then If w j {\displaystyle w_{j}} and x i {\displaystyle x_{i}} have the same class, and w k {\displaystyle w_{k}} and x i {\displaystyle x_{i}} have different classes, then w j ← w j + α t ( x i − w j ) {\displaystyle w_{j}\leftarrow w_{j}+\alpha _{t}(x_{i}-w_{j})} and w k ← w k − α t ( x i − w k ) {\displaystyle w_{k}\leftarrow w_{k}-\alpha _{t}(x_{i}-w_{k})} . If w k {\displaystyle w_{k}} and x i {\displaystyle x_{i}} have the same class, and w j {\displaystyle w_{j}} and x i {\displaystyle x_{i}} have different classes, then w j ← w j − α t ( x i − w j ) {\displaystyle w_{j}\leftarrow w_{j}-\alpha _{t}(x_{i}-w_{j})} and w k ← w k + α t ( x i − w k ) {\displaystyle w_{k}\leftarrow w_{k}+\alpha _{t}(x_{i}-w_{k})} . If w j {\displaystyle w_{j}} and w k {\displaystyle w_{k}} and x i {\displaystyle x_{i}} have the same class, then w j ← w j − ϵ α t ( x i − w j ) {\displaystyle w_{j}\leftarrow w_{j}-\epsilon \alpha _{t}(x_{i}-w_{j})} and w k ← w k + ϵ α t ( x i − w k ) {\displaystyle w_{k}\leftarrow w_{k}+\epsilon \alpha _{t}(x_{i}-w_{k})} . If w k {\displaystyle w_{k}} and x i {\displaystyle x_{i}} have different classes, and w j {\displaystyle w_{j}} and x i {\displaystyle x_{i}} have different classes, then the original paper simply does not explain what happens in this case, but presumably nothing happens in this case. Otherwise, skip. Note that condition min ( d j d k , d k d j ) > s {\displaystyle \min \left({\frac {d_{j}}{d_{k}}},{\frac {d_{k}}{d_{j}}}\right)>s} , where s = 1 − w 1 + w {\displaystyle s={\frac {1-w}{1+w}}} , precisely means that the point x i {\displaystyle x_{i}} falls between two Apollonian spheres.

  Read more →
• 

  Simulation decomposition

  SimDec, or Simulation decomposition, is a hybrid uncertainty and sensitivity analysis method, for visually examining the relationships between the output and input variables of a computational model. SimDec maps multivariable scenarios onto the distribution of the model output. This visual analytics approach exposes the underlying nature of the model behavior, including its nonlinear and multivariate interaction effects. SimDec can be used in any range of science, engineering, and social domains. Existing applications include business and environmental issues. == Method == SimDec operates on Monte Carlo simulation (or measured) data where both output and input values are recorded. At least one thousand observations (or simulated iterations) are typically recommended to preserve the readability of the resulting histograms. An outline of the decomposition algorithm, which is readily available in multiple programming languages, proceeds as follows: Select the input variables for decomposition. One can use sensitivity indices (see variance-based sensitivity analysis) to define the most influential variables for decomposition or choose them manually according to the decision-problem context (for example, only those input variables that the decision-maker can act upon). Two to three input variables, ordered by decreasing value of their sensitivity indices, usually provide the most meaningful decomposition results. Divide the inputs into states. The numeric ranges of the inputs are split into several intervals with an equal number of observations in each. For categorical variables, the categories represent states. Form scenarios. All combinations of states of the selected input variables produce unique scenarios or subsets of the data. For example, if the range of X2 is divided into low, medium and high, and X3 takes values of 1 or 2, six scenarios are formed: (i) X2 low & X3 = 1, (ii) X2 low & X3 = 2, (iii) X2 medium & X3 = 1, (iv) X2 medium & X3 = 2, (v) X2 high & X3 = 1, and (vi) X2 high & X3 = 2. Assign scenarios to each output value. The simulation data is used to define the scenario index for each simulation run. For example, if an X2 value falls into the low state and X3 is equal to 2, the corresponding scenario, defined in Step 3, is (ii). Color-code the output distribution. When all output values are assigned scenario indices, they are plotted as series in a stacked histogram, visually separated by color-coding. For ease of visual perception, the states of the most influential input variable are assigned distinct colors, and all the remaining partitions take shades of those colors (see Figure). All of these steps can be run automatically on the given data using the open-source SimDec packages currently available in Python, R, Julia, and Matlab. A SimDec template in Excel runs a Monte Carlo simulation of a spreadsheet model but possesses only a manual option for input selection. == How to read SimDec == === Histogram === Histogram is an approximate representation of the distribution of numerical data. Its horizontal axis shows the range of the variable of interest, and its vertical axis denotes count, also called frequency, or, if divided by the total number of data points, probability. The distribution alone can supply only limited information about the data – its minimum, maximum, and shape (where the most of data occurs). === Judging the importance of inputs === If an input variable has no effect on the output, its states (e.g., low & high) would lie on top of each other on the SimDec histogram, occupying fully overlapping ranges of the output. If an input variable has a strong effect and explains most of the variance of the output, the border between its states on the SimDec histogram would be vertical. Such visualization has an important decision-making implication – e.g., if the high state of X can be achieved, it would guarantee a certain range of Y. All cases in-between with low-to-strong effects would show a diagonal border between the states. The less they overlap, the larger the effect of X on Y. While the horizontal displacement of sub-distributions on the SimDec histogram is the key to interpreting the results, the vertical disposition of sub-distributions is just a technical matter of the order of plotting the series of the stacked histogram. === Exploring the interaction of inputs === When two or more input variables are used for decomposition, it becomes possible to examine their joint effects. A schematic visualization portrays how different types of joint effects of input variables on the output appear on SimDec visualization. Understanding the nature of interaction effects in a computational model and its behavior in general is crucial for effective decision-making. == Limitations == The SimDec method has several limitations: It is based on Monte Carlo simulation and thus requires running a computational model a thousand of times or more. To models that take hours to evaluate once, it would be impossible to use SimDec (unless a supercomputer and/or large of time are available). SimDec is based on a histogram, thus, for binary or categorical output variables, the visualization would be very limited (e.g., only a few bins). The more input variables one selects for the decomposition, the less readable the histogram becomes. Only cases with two and three input variables are presented in.

  Read more →
• 

  Endomondo

  Endomondo is a health and wellness website. It allows users to track their health statistics and provides insights on fitness trends. Originally launched in 2007, Endomondo was acquired by Under Armour in 2015. Under Armour shut down Endomondo in 2020, but, by 2024, Endomondo re-launched as its own entity. == History == Endomondo started in Denmark in 2007 by Mette Lykke, Christian Birk and Jakob Nordenhof Jønck. In 2011, the company opened an office in Silicon Valley, USA, but kept its research and development department in Denmark. In 2013, Endomondo LLC was listed in Red Herring as a European finalists for promising start-ups. The same year, Christian Birk and Jakob Nordenhof Jønck left the daily operation of the company, but kept co-ownership. In February 2015, Endomondo LLC was acquired by athletic apparel maker Under Armour for $85 million. Endomondo, at that time, had over 20 million users. In October 2020, Under Armour announced that Endomondo would be shutting down and selling off MyFitnessPal to the private equity firm Francisco Partners for $345 million. Service stopped on 31 December 2020, giving customers until 15 February 2021 to download an archive of their historic data. In 2024, Endomondo.com was brought back online as a professional fitness guidance website. == Features == Endomondo provides numerous workouts, guidance on exercises, performance-enhancing nutrition, and tips. Previously, Endomondo was able to track numerous fitness attributes such as running routes, distance, duration, and calories. The software helped analyze performance and recommend improvements. There was a free and a paid version available of Endomondo. The free version had advertisements. The paid Premium version was free of advertisements and included additional features such as the possibility to create one's own training plan. The offering of additional features was different between the Android, IOS and Windows platforms, and had significantly better features for tracking performance over time than UnderArmours suggested replacement. Endomondo offered challenges of various types to the user and allowed users to create their own challenges.

  Read more →
• 

  Dendral

  Dendral was a project in artificial intelligence (AI) of the 1960s, and the computer software expert system that it produced. Its primary aim was to study hypothesis formation and discovery in science. For that, a specific task in science was chosen: help organic chemists in identifying unknown organic molecules, by analyzing their mass spectra and using knowledge of chemistry. It was done at Stanford University by Edward Feigenbaum, Bruce G. Buchanan, Joshua Lederberg, and Carl Djerassi, along with a team of highly creative research associates and students. It began in 1964 and spans approximately half the history of AI research. The software program Dendral is considered the first expert system because it automated the decision-making process and problem-solving behavior of organic chemists. The project consisted of research on two main programs Heuristic Dendral and Meta-Dendral, and several sub-programs. It was written in the Lisp programming language, which was considered the language of AI because of its flexibility. Many systems were derived from Dendral, including MYCIN, MOLGEN, PROSPECTOR, XCON, and STEAMER. There are many other programs today for solving the mass spectrometry inverse problem, see List of mass spectrometry software, but they are no longer described as 'artificial intelligence', just as structure searchers. The name Dendral is an acronym of the term "Dendritic Algorithm". == Heuristic Dendral == Heuristic Dendral is a program that uses mass spectra or other experimental data together with a knowledge base of chemistry to produce a set of possible chemical structures that may be responsible for producing the data. A mass spectrum of a compound is produced by a mass spectrometer, and is used to determine its molecular weight, the sum of the masses of its atomic constituents. For example, the compound water (H2O), has a molecular weight of 18 since hydrogen has a mass of 1.01 and oxygen 16.00, and its mass spectrum has a peak at 18 units. Heuristic Dendral would use this input mass and the knowledge of atomic mass numbers and valence rules, to determine the possible combinations of atomic constituents whose mass would add up to 18. As the weight increases and the molecules become more complex, the number of possible compounds increases drastically. Thus, a program that is able to reduce this number of candidate solutions through the process of hypothesis formation is essential. New graph-theoretic algorithms were invented by Lederberg, Harold Brown, and others that generate all graphs with a specified set of nodes and connection-types (chemical atoms and bonds) -- with or without cycles. Moreover, the team was able to prove mathematically that the generator is complete, in that it produces all graphs with the specified nodes and edges, and that it is non-redundant, in that the output contains no equivalent graphs (e.g., mirror images). The CONGEN program, as it became known, was developed largely by computational chemists Ray Carhart, Jim Nourse, and Dennis Smith. It was useful to chemists as a stand-alone program to generate chemical graphs showing a complete list of structures that satisfy the constraints specified by a user. == Meta-Dendral == Meta-Dendral is a machine learning system that receives the set of possible chemical structures and corresponding mass spectra as input, and proposes a set of rules of mass spectrometry that correlate structural features with processes that produce the mass spectrum. These rules would be fed back to Heuristic Dendral (in the planning and testing programs described below) to test their applicability. Thus, "Heuristic Dendral is a performance system and Meta-Dendral is a learning system". The program is based on two important features: the plan-generate-test paradigm and knowledge engineering. === Plan-generate-test paradigm === The plan-generate-test paradigm is the basic organization of the problem-solving method, and is a common paradigm used by both Heuristic Dendral and Meta-Dendral systems. The generator (later named CONGEN) generates potential solutions for a particular problem, which are then expressed as chemical graphs in Dendral. However, this is feasible only when the number of candidate solutions is minimal. When there are large numbers of possible solutions, Dendral has to find a way to put constraints that rules out large sets of candidate solutions. This is the primary aim of Dendral planner, which is a “hypothesis-formation” program that employs “task-specific knowledge to find constraints for the generator”. Last but not least, the tester analyzes each proposed candidate solution and discards those that fail to fulfill certain criteria. This mechanism of plan-generate-test paradigm is what holds Dendral together. === Knowledge Engineering === The primary aim of knowledge engineering is to attain a productive interaction between the available knowledge base and problem solving techniques. This is possible through development of a procedure in which large amounts of task-specific information is encoded into heuristic programs. Thus, the first essential component of knowledge engineering is a large “knowledge base.” Dendral has specific knowledge about the mass spectrometry technique, a large amount of information that forms the basis of chemistry and graph theory, and information that might be helpful in finding the solution of a particular chemical structure elucidation problem. This “knowledge base” is used both to search for possible chemical structures that match the input data, and to learn new “general rules” that help prune searches. The benefit Dendral provides the end user, even a non-expert, is a minimized set of possible solutions to check manually. == Heuristics == A heuristic is a rule of thumb, an algorithm that does not guarantee a solution, but reduces the number of possible solutions by discarding unlikely and irrelevant solutions. The use of heuristics to solve problems is called "heuristics programming", and was used in Dendral to allow it to replicate in machines the process through which human experts induce the solution to problems via rules of thumb and specific information. Heuristics programming was a major approach and a giant step forward in artificial intelligence, as it allowed scientists to finally automate certain traits of human intelligence. It became prominent among scientists in the late 1940s through George Polya’s book, How to Solve It: A New Aspect of Mathematical Method. As Herbert A. Simon said in The Sciences of the Artificial, "if you take a heuristic conclusion as certain, you may be fooled and disappointed; but if you neglect heuristic conclusions altogether you will make no progress at all." == History == During the mid 20th century, the question "can machines think?" became intriguing and popular among scientists, primarily to add humanistic characteristics to machine behavior. John McCarthy, who was one of the prime researchers of this field, termed this concept of machine intelligence as "artificial intelligence" (AI) during the Dartmouth summer in 1956. AI is usually defined as the capacity of a machine to perform operations that are analogous to human cognitive capabilities. Much research to create AI was done during the 20th century. Also around the mid 20th century, science, especially biology, faced a fast-increasing need to develop a "man-computer symbiosis", to aid scientists in solving problems. For example, the structural analysis of myoglobin, hemoglobin, and other proteins relentlessly needed instrumentation development due to its complexity. In the early 1960s, Joshua Lederberg started working with computers and quickly became tremendously interested in creating interactive computers to help him in his exobiology research. Specifically, he was interested in designing computing systems to help him study alien organic compounds. Lederberg had been heading a team designing instruments for the Mars Viking lander to search for precursor molecules of life in samples of the Mars surface, using a mass spectrometer coupled with a minicomputer. As he was not an expert in either chemistry or computer programming, he collaborated with Stanford chemist Carl Djerassi to help him with chemistry, and Edward Feigenbaum with programming, to automate the process of determining chemical structures from raw mass spectrometry data. Feigenbaum was an expert in programming languages and heuristics, and helped Lederberg design a system that replicated the way Djerassi solved structure elucidation problems. They devised a system called Dendritic Algorithm (Dendral) that was able to generate possible chemical structures corresponding to the mass spectrometry data as an output. Dendral then was still very inaccurate in assessing spectra of ketones, alcohols, and isomers of chemical compounds. Thus, Djerassi "taught" general rules to Dendral that could help eliminate most of the "chemically implausible" structures, and p

  Read more →
• 

  Andrej Karpathy

  Andrej Karpathy (born 23 October 1986) is a Slovak-Canadian AI researcher, who co-founded and formerly worked at OpenAI, where he specialized in deep learning and computer vision. He also worked as the director of artificial intelligence and Autopilot Vision at Tesla, and in 2024 he founded Eureka Labs, an AI education platform. In 2026 he joined Anthropic as part of the pretraining team. == Education and early life == Karpathy was born in Bratislava, Czechoslovakia (now Slovakia), and moved with his family to Toronto when he was 15. He completed his Computer Science and Physics bachelor's degrees at University of Toronto in 2009 and his master's degree at University of British Columbia in 2011, where he worked on physically simulated figures (for example, a simulated runner or a simulated person in a crowd) with his adviser Michiel van de Panne. In 2006, Karpathy began posting videos on YouTube on his channel, badmephisto. He garnered fame by posting Rubik's cube tutorials which have been used by famous speedcubers such as Feliks Zemdegs. The channel has over 9 million views as of June 2025. Karpathy received a PhD from Stanford University in 2015 under the supervision of Fei-Fei Li, focusing on the intersection of natural language processing and computer vision, and deep learning models suited for this task. == Career and research == He authored and was the primary instructor of the first deep learning course at Stanford, CS 231n: Convolutional Neural Networks for Visual Recognition. The course became one of the largest classes at Stanford, growing from 150 students in 2015 to 750 in 2017. Karpathy is a founding member of the artificial intelligence research group OpenAI, where he was a research scientist from 2015 to 2017. In June 2017 he became Tesla's director of artificial intelligence and reported to Elon Musk. He was named one of MIT Technology Review's Innovators Under 35 for 2020. After taking a several-months-long sabbatical from Tesla, he announced he was leaving the company in July 2022. As of February 2023, he makes YouTube videos on how to create artificial neural networks. On February 9, 2023, Karpathy announced he was returning to OpenAI. A year later on February 13, 2024, an OpenAI spokesperson confirmed that Karpathy had left OpenAI. In the same year, he was named one of Time Magazine's 100 Most Influential People in AI. On July 16, 2024, Karpathy announced on his X account that he started a new AI education company called Eureka Labs. Their first product was the AI course, LLM101n. He also has a broader educational effort, the "Zero to Hero" series on LLM fundamentals. The company also advocates for AI teaching assistants, a concept which has been criticized due to data privacy concerns and the removal of personal connection between teacher and student. In February 2025, Karpathy coined the term vibe coding to describe how AI tools allow hobbyists to construct apps and websites just by typing prompts. On May 19, 2026, he announced that he joined Anthropic via a statement on X, while the company stated that he will be leading a team for research in pretraining.

  Read more →
• 

  Business rule management system

  A BRMS or business rule management system is a software system used to define, deploy, execute, monitor and maintain the variety and complexity of decision logic that is used by operational systems within an organization or enterprise. This logic, also referred to as business rules, includes policies, requirements, and conditional statements that are used to determine the tactical actions that take place in applications and systems. == Overview == A BRMS includes, at minimum: A repository, allowing decision logic to be externalized from core application code Tools, allowing both technical developers and business experts to define and manage decision logic A runtime environment, allowing applications to invoke decision logic managed within the BRMS and execute it using a business rules engine The top benefits of a BRMS include: Reduced or removed reliance on IT departments for changes in live systems. Although, QA and Rules testing would still be needed in any enterprise system. Increased control over implemented decision logic for compliance and better business management including audit logs, impact simulation and edit controls. The ability to express decision logic with increased precision, using a business vocabulary syntax and graphical rule representations (decision tables, decision models, trees, scorecards and flows) Improved efficiency of processes through increased decision automation. Some disadvantages of the BRMS include: Extensive subject matter expertise can be required for vendor specific products. In addition to appropriate design practices (such as Decision Modeling), technical developers must know how to write rules and integrate software with existing systems Poor rule harvesting approaches can lead to long development cycles, though this can be mitigated with modern approaches like the Decision Model and Notation (DMN) standard. Integration with existing systems is still required and a BRMS may add additional security constraints. Reduced IT department reliance may never be a reality due to continued introduction to new business rule considerations or object model perturbations The coupling of a BRMS vendor application to the business application may be too tight to replace with another BRMS vendor application. This can lead to cost to benefits issues. The emergence of the DMN standard has mitigated this to some degree. Most BRMS vendors have evolved from rule engine vendors to provide business-usable software development lifecycle solutions, based on declarative definitions of business rules executed in their own rule engine. BRMSs are increasingly evolving into broader digital decisioning platforms that also incorporate decision intelligence and machine learning capabilities. However, some vendors come from a different approach (for example, they map decision trees or graphs to executable code). Rules in the repository are generally mapped to decision services that are naturally fully compliant with the latest SOA, Web Services, or other software architecture trends. == Related software approaches == In a BRMS, a representation of business rules maps to a software system for execution. A BRMS therefore relates to model-driven engineering, such as the model-driven architecture (MDA) of the Object Management Group (OMG). It is no coincidence that many of the related standards come under the OMG banner. A BRMS is a critical component for Enterprise Decision Management as it allows for the transparent and agile management of the decision-making logic required in systems developed using this approach. == Associated standards == The OMG Decision Model and Notation standard is designed to standardize elements of business rules development, specially decision table representations. There is also a standard for a Java Runtime API for rule engines JSR-94. OMG Business Motivation Model (BMM): A model of how strategies, processes, rules, etc. fit together for business modeling OMG SBVR: Targets business constraints as opposed to automating business behavior OMG Production Rule Representation (PRR): Represents rules for production rule systems that make up most BRMS' execution targets OMG Decision Model and Notation (DMN): Represents models of decisions, which are typically managed by a BRMS RuleML provides a family of rule mark-up languages that could be used in a BRMS and with W3C RIF it provides a family of related rule languages for rule interchange in the W3C Semantic Web stack Many standards, such as domain-specific languages, define their own representation of rules, requiring translations to generic rule engines or their own custom engines. Other domains, such as PMML, also define rules.

  Read more →
• 

  Index locking

  In databases an index is a data structure, part of the database, used by a database system to efficiently navigate access to user data. Index data are system data distinct from user data, and consist primarily of pointers. Changes in a database (by insert, delete, or modify operations), may require indexes to be updated to maintain accurate user data accesses. Index locking is a technique used to maintain index integrity. A portion of an index is locked during a database transaction when this portion is being accessed by the transaction as a result of attempt to access related user data. Additionally, special database system transactions (not user-invoked transactions) may be invoked to maintain and modify an index, as part of a system's self-maintenance activities. When a portion of an index is locked by a transaction, other transactions may be blocked from accessing this index portion (blocked from modifying, and even from reading it, depending on lock type and needed operation). Index Locking Protocol guarantees that phantom read phenomenon won't occur. Index locking protocol states: Every relation must have at least one index. A transaction can access tuples only after finding them through one or more indices on the relation A transaction Ti that performs a lookup must lock all the index leaf nodes that it accesses, in S-mode, even if the leaf node does not contain any tuple satisfying the index lookup (e.g. for a range query, no tuple in a leaf is in the range) A transaction Ti that inserts, updates or deletes a tuple ti in a relation r must update all indices to r and it must obtain exclusive locks on all index leaf nodes affected by the insert/update/delete The rules of the two-phase locking protocol must be observed. Specialized concurrency control techniques exist for accessing indexes. These techniques depend on the index type, and take advantage of its structure. They are typically much more effective than applying to indexes common concurrency control methods applied to user data. Notable and widely researched are specialized techniques for B-trees (B-Tree concurrency control) which are regularly used as database indexes. Index locks are used to coordinate threads accessing indexes concurrently, and typically shorter-lived than the common transaction locks on user data. In professional literature, they are often called latches.

  Read more →
• 

  Chainer

  Chainer is an open source deep learning framework written purely in Python on top of NumPy and CuPy Python libraries. The development is led by Japanese venture company Preferred Networks in partnership with IBM, Intel, Microsoft, and Nvidia. Chainer is notable for its early adoption of "define-by-run" scheme, as well as its performance on large scale systems. The first version was released in June 2015 and has gained large popularity in Japan since then. Furthermore, in 2017, it was listed by KDnuggets in top 10 open source machine learning Python projects. In December 2019, Preferred Networks announced the transition of its development effort from Chainer to PyTorch and it will only provide maintenance patches after releasing v7. == Define-by-run == Chainer was the first deep learning framework to introduce the define-by-run approach. The traditional procedure to train a network was in two phases: define the fixed connections between mathematical operations (such as matrix multiplication and nonlinear activations) in the network, and then run the actual training calculation. This is called the define-and-run or static-graph approach. Theano and TensorFlow are among the notable frameworks that took this approach. In contrast, in the define-by-run or dynamic-graph approach, the connection in a network is not determined when the training is started. The network is determined during the training as the actual calculation is performed. One of the advantages of this approach is that it is intuitive and flexible. If the network has complicated control flows such as conditionals and loops, in the define-and-run approach, specially designed operations for such constructs are needed. On the other hand, in the define-by-run approach, programming language's native constructs such as if statements and for loops can be used to describe such flow. This flexibility is especially useful to implement recurrent neural networks. Another advantage is ease of debugging. In the define-and-run approach, if an error (such as numeric error) has occurred in the training calculation, it is often difficult to inspect the fault, because the code written to define the network and the actual place of the error are separated. In the define-by-run approach, you can just suspend the calculation with the language's built-in debugger and inspect the data that flows on your code of the network. Define-by-run has gained popularity since the introduction by Chainer and is now implemented in many other frameworks, including PyTorch and TensorFlow. == Extension libraries == Chainer has four extension libraries, ChainerMN, ChainerRL, ChainerCV and ChainerUI. ChainerMN enables Chainer to be used on multiple GPUs with performance significantly faster than other deep learning frameworks. A supercomputer running Chainer on 1024 GPUs processed 90 epochs of ImageNet dataset on ResNet-50 network in 15 minutes, which is four times faster than the previous record held by Facebook. ChainerRL adds state of art deep reinforcement learning algorithms, and ChainerUI is a management and visualization tool. == Applications == Chainer is used as the framework for PaintsChainer, a service which does automatic colorization of black and white, line only, draft drawings with minimal user input.

  Read more →
• 

  Knowledge processing for robots

  KnowRob (Knowledge processing for robots) is a system which combines knowledge representation and reasoning methods to acquire and ground knowledge. This system is the backbone of openEASE. both under developing at the Institute for Artificial Intelligence at the University of Bremen, Germany. == The framework == KnowRob can serve as a common sense framework for the integration of knowledge. This knowledge can be static encyclopedic knowledge, common sense knowledge, task descriptions, environment models, object information, observed actions, etc., which can come from different sources, like manually axiomatized, derived from observations, or imported from the web. KnowRob has been used by different research groups, as the Rice University using the ontological knowledge base in a robotic platform. As well by the Eindhoven University of Technology research group competing in the RoboCup league, in the "at Home" category, with the RoboEarth project. As well, KnowRob is mentioned in the work of some research groups from the Lucian Blaga University of Sibiu, Middle East Technical University in their combination of different knowledge bases, Keio University as related work because of the ontology service, University of Texas at Austin as related work as well because of the relation with the work presented, Hanyang University as related work as an OWL based knowledge processing framework. == Representations == To represent the knowledge, KnowRob uses the OWL ontology language and an extended first-order logic knowledge representation with computable predicates. To give the order of subactions, KnowRob includes a pair-wise ordering constrain, which gives a partial ordering. KnowRob adopts the closed-world assumption Prolog, and an open-world assumption by the use of computables. To include reasoning rules into Prolog, KnowRob uses an inference procedure beyond the capabilities of OWL to extract information about tasks executions. In its second version, KnowRob provides a logic interface to the hybrid reasoning kernel as a logic based language. This language presents the hybrid reasoning kernel as if everything were entities retrievable by providing partial descriptions for them. This entities descriptions include objects, their parts, and articulation models, environments composed of objects, software components, actions, and events. === Episodic memories === Episodic memory is related to the experience information, which is organized temporally and spatially, alongside combined with context information. In KnowRob, an episodic memory is understood as a recording that the agent makes of the ongoing activity, which includes very detailed information about the actions, motions, their purposes, effects and the behavior they generate, it also includes the images captured during execution, etc. == Usage == The knowledge is computed by external methods using Prolog queries. In the second version of the KnowRob system, is included a better structure of the packages and documentations. Which includes some extensions from the previous version, as well as a logic based language. For example, a cup description from perception can be represented in this language as: entity(Cup,[an, object, [type, cup], [shape, cylinder], [color, orange]]) As well, a controller could represent the same object as: entity(Cup, [an, object, [type, cup], [proper_physical_parts, [an, object, [type, handle], [grasp−pose, G−pose]]]]) The interface language is comparable to other query languages for symbolic knowledge bases. KnowRob's query language integrates reasoning methods, such as the simulation-based reasoning. == Goals == The goal of the KnowRob framework is to make semantic knowledge available for service robots. It is able to answer queries about missing information in vague instructions for tasks. This is possible with the actions hierarchical representation and information about objects which can be included in certain action.

  Read more →
• 

  Innovation Center for Artificial Intelligence

  The Innovation Center for Artificial Intelligence (ICAI) is a Dutch national network focused on joint technology development between academia, industry and government in the area of artificial intelligence (AI). The initiative was launched in April 2018 and is based at Amsterdam Science Park. As of 2024, the director of the ICAI is Maarten de Rijke. In November 2018, ICAI announced its contribution to AINED, the first iteration of the Dutch National AI Strategy. In January 2023, Maastricht University announced the ROBUST program, led by the Innovation Center for Artificial Intelligence (ICAI) and supported by the University of Amsterdam and others. This initiative focuses on advancing research in trustworthy AI technology across various sectors, notably healthcare and energy, in the Netherlands. The program's plan includes the creation of 17 new labs and the appointment of PhD candidates, backed by a €25 million funding from the Dutch Research Council (NWO). == Labs == The ICAI network is linked to several collaborative labs: Thira Lab (Imaging): Thirona, Delft Imaging Systems and Radboud UMC, founded March 2019 AIMLab (AI for Medical Imaging): Uva and Inception Institute of Artificial Intelligence from the United Arab Emirates, founded March 2019 AFL (AI for Fintech): ING and Delft University of Technology, founded March 2019 Police Lab AI: Dutch National Police, founded January 2019 Elsevier AI Lab: Uva and Elsevier, founded October 2018 AIRLab Delft (AI for Retail Robotics): TU Delft Robotics and AholdDelhaize, founded November 2018 Quva Lab (Deep Vision): Uva and Qualcomm, founded 2016 (prior to ICAI) AIRLab Amsterdam (AI for Retail): Uva and AholdDelhaize, founded April 2018 DeltaLab (Deep Learning Technologies Amsterdam): Uva and Bosch, founded April 2017 (prior to ICAI) AI4SE (AI for Software Engineering Lab) Delft University of Technology and JetBrains, founded October 2023 Atlas Lab: Uva and TomTom (TOM2)

  Read more →
• 

  Chatbot

  A chatbot (originally chatterbot) is a software application or web interface designed to converse through text or speech. Modern chatbots are typically online and use generative artificial intelligence systems that are capable of maintaining a conversation with a user in natural language and simulating the way a human would behave as a conversational partner. Such chatbots often use deep learning and natural language processing. Simpler chatbots have existed for decades. Chatbots have gained popularity during the AI boom of the 2020s, with the releases of generative AI chatbots such as ChatGPT, Gemini, Claude, and Grok. These chatbots typically use fine-tuned large language models to generate text. A major area where chatbots have long been used is customer service and support, with various sorts of virtual assistants. == History == === Turing test === In 1950, Alan Turing published an article entitled "Computing Machinery and Intelligence" in which he proposed what is now called the Turing test as a criterion of intelligence. This criterion depends on the ability of a computer program to impersonate a human in a real-time written conversation with a human judge, to the extent that the judge is incapable of reliably distinguishing, on the basis of the conversational content alone, between the program and a real human. === Early chatbots === Joseph Weizenbaum's program ELIZA was first published in 1966. Weizenbaum did not claim that ELIZA was genuinely intelligent, and the introduction to his paper presented it more as a debunking exercise:In artificial intelligence, machines are made to behave in wondrous ways, often sufficient to dazzle even the most experienced observer. But once a particular program is unmasked, once its inner workings are explained, its magic crumbles away; it stands revealed as a mere collection of procedures. The observer says to himself "I could have written that". With that thought, he moves the program in question from the shelf marked "intelligent", to that reserved for curios. The object of this paper is to cause just such a re-evaluation of the program about to be "explained". Few programs ever needed it more. ELIZA's key method of operation involves the recognition of clue words or phrases in the input, and the output of the corresponding pre-prepared or pre-programmed responses that can move the conversation forward in an apparently meaningful way (e.g. by responding to any input that contains the word 'MOTHER' with 'TELL ME MORE ABOUT YOUR FAMILY'). Thus an illusion of understanding is generated, even though the processing involved has been merely superficial. ELIZA showed that such an illusion is surprisingly easy to generate because human judges are ready to give the benefit of the doubt when conversational responses are capable of being interpreted as "intelligent". Following ELIZA, psychiatrist Kenneth Colby developed PARRY in 1972. From 1978 to some time after 1983, the CYRUS project led by Janet Kolodner constructed a chatbot simulating Cyrus Vance (57th United States Secretary of State). It used case-based reasoning, and updated its database daily by parsing wire news from United Press International. The program was unable to process the news items subsequent to the surprise resignation of Cyrus Vance in April 1980, and the team constructed another chatbot simulating his successor, Edmund Muskie. In 1984, an interactive version of the program Racter was released which acted as a chatbot. A.L.I.C.E. was released in 1995. This uses a markup language called AIML, which is specific to its function as a conversational agent, and has since been adopted by various other developers of, so-called, Alicebots. A.L.I.C.E. is a weak AI without any reasoning capabilities. It is based on a similar pattern matching technique as ELIZA in 1966. This is not strong AI, which would require sapience and logical reasoning abilities. Jabberwacky, released in 1997, learns new responses and context based on real-time user interactions, rather than being driven from a static database. Chatbot competitions focus on the Turing test or more specific goals. Two such annual contests are the Loebner Prize and The Chatterbox Challenge (the latter has been offline since 2015, however, materials can still be found from web archives). Pre-dating the current generation of large language models, Gavagai, a Swedish language technology startup, created a Twitter-based bot in 2015 and DBpedia created a chatbot during the 2017 Google Summer of Code that communicated through Facebook Messenger. === Modern chatbots based on large language models === Modern chatbots like ChatGPT are often based on foundational large language models called generative pre-trained transformers (GPT). They are based on a deep learning architecture called the transformer, which contains artificial neural networks. They generate text after being trained on a large text corpus, and have emergent abilities that they are not specifically trained for. Chatbots integrated into apps and websites can call image-generation models or search the web. Some platforms also enable users to interact with conversational interfaces directly through web-based chat environments, allowing real-time assistance, content generation, and task automation without requiring software installation. == Application == === Messaging apps === Many companies' chatbots run on messaging apps or simply via SMS. They are used for B2C customer service, sales and marketing. In 2016, Facebook Messenger allowed developers to place chatbots on their platform. There were 30,000 bots created for Messenger in the first six months, rising to 100,000 by September 2017. Since September 2017, this has also been as part of a pilot program on WhatsApp. Airlines KLM and Aeroméxico both announced their participation in the testing; both airlines had previously launched customer services on the Facebook Messenger platform. The bots usually appear as one of the user's contacts, but can sometimes act as participants in a group chat. Many banks, insurers, media companies, e-commerce companies, airlines, hotel chains, retailers, health care providers, government entities, and restaurant chains have used chatbots to answer simple questions, increase customer engagement, for promotion, and to offer additional ways to order from them. Chatbots are also used in market research to collect short survey responses. A 2017 study showed 4% of companies used chatbots. In a 2016 study, 80% of businesses said they intended to have one by 2020. ==== As part of company apps and websites ==== Previous generations of chatbots were present on company websites, e.g. Ask Jenn from Alaska Airlines which debuted in 2008 or Expedia's virtual customer service agent which launched in 2011. The newer generation of chatbots includes IBM Watson-powered "Rocky", introduced in February 2017 by the New York City-based e-commerce company Rare Carat to provide information to prospective diamond buyers. ==== Chatbot sequences ==== Used by marketers to script sequences of messages, very similar to an autoresponder sequence. Such sequences can be triggered by user opt-in or the use of keywords within user interactions. After a trigger occurs a sequence of messages is delivered until the next anticipated user response. Each user response is used in the decision tree to help the chatbot navigate the response sequences to deliver the correct response message. === Company internal platforms === Companies have used chatbots for customer support, human resources, or in Internet-of-Things (IoT) projects. Overstock.com, for one, has reportedly launched a chatbot named Mila to attempt to automate certain processes when customer service employees request sick leave. Other large companies such as Lloyds Banking Group, Royal Bank of Scotland, Renault and Citroën are now using chatbots instead of call centres with humans to provide a first point of contact. In large companies, like in hospitals and aviation organizations, chatbots are also used to share information within organizations, and to assist and replace service desks. === Customer service === Chatbots have been proposed as a replacement for customer service departments. In 2026, The Financial Times reported on agentic chatbots that could do shopping for customers once given instructions. In 2016, Russia-based Tochka Bank launched a chatbot on Facebook for a range of financial services, including a possibility of making payments. In July 2016, Barclays Africa also launched a Facebook chatbot. === Healthcare === Chatbots are also appearing in the healthcare industry. A study suggested that physicians in the United States believed that chatbots would be most beneficial for scheduling doctor appointments, locating health clinics, or providing medication information. A 2025 review found that participants often rated chatbot responses as more empathic than those from clinicians. In 2020, WhatsApp worked with th

  Read more →
• 

  Goal node (computer science)

  In computer science, a goal node is a node in a graph that meets defined criteria for success or termination. Heuristical artificial intelligence algorithms, like A and B, attempt to reach such nodes in optimal time by defining the distance to the goal node. When the goal node is reached, A defines the distance to the goal node as 0 and all other nodes' distances as positive values.

  Read more →
• 

  Open Neural Network Exchange

  The Open Neural Network Exchange (ONNX) [ˈɒnɪks] is an open-source artificial intelligence ecosystem of technology companies and research organizations that establish open standards for representing machine learning algorithms and software tools to enable a standard format for representing machine learning models. ONNX is available on GitHub. == History == ONNX was originally named Toffee and was developed by the PyTorch team at Facebook. In September 2017 it was renamed to ONNX and announced by Facebook and Microsoft. Later, IBM, Huawei, Intel, AMD, Arm and Qualcomm announced support for the initiative. In October 2017, Microsoft announced that it would add its Cognitive Toolkit and Project Brainwave platform to the initiative. In November 2019 ONNX was accepted as graduate project in Linux Foundation AI. In October 2020 Zetane Systems became a member of the ONNX ecosystem. == Intent == The initiative targets: === Framework interoperability === Enable developers to move machine learning models between different frameworks, which may be used at different stages of the development process, such as training, architecture design, or deployment on mobile devices. === Shared optimization === Provide a common representation that can be used by hardware vendors and other developers to apply optimizations to artificial neural network models across multiple machine learning frameworks. == Contents == ONNX provides definitions of an extensible computation graph model, built-in operators and standard data types, focused on inferencing (evaluation).. The container format is Protocol Buffers. Each computation dataflow graph is a list of nodes that form an acyclic graph. Nodes have inputs and outputs. Each node is a call to an operator. Metadata documents the graph. Built-in operators are to be available on each ONNX-supporting framework. ONNX models can be trained in a single framework, such as PyTorch or TensorFlow, and then exported to ONNX. This format allows models to be transferred from the training framework to other environments for testing or deployment. Once a model is in ONNX format, it can be executed in different runtime systems or on various hardware platforms, such as GPUs or specialized AI accelerators. Using a common format enables the same model representation to be used across multiple systems and frameworks.

  Read more →