AI Face Swap Gif

AI Face Swap Gif — independent reviews, comparisons, pricing and step-by-step guides on Aizhi.

• 

  Situated approach (artificial intelligence)

  In artificial intelligence research, the situated approach builds agents that are designed to behave effectively successfully in their environment. This requires designing AI "from the bottom-up" by focussing on the basic perceptual and motor skills required to survive. The situated approach gives a much lower priority to abstract reasoning or problem-solving skills. The approach was originally proposed as an alternative to traditional approaches (that is, approaches popular before 1985 or so). After several decades, classical AI technologies started to face intractable issues (e.g. combinatorial explosion) when confronted with real-world modeling problems. All approaches to address these issues focus on modeling intelligences situated in an environment. They have become known as the situated approach to AI. == Emergence of a concept == === From traditional AI to Nouvelle AI === During the late 1980s, the approach now known as Nouvelle AI (Nouvelle means new in French) was pioneered at the MIT Artificial Intelligence Laboratory by Rodney Brooks. As opposed to classical or traditional artificial intelligence, Nouvelle AI purposely avoided the traditional goal of modeling human-level performance, but rather tries to create systems with intelligence at the level of insects, closer to real-world robots. But eventually, at least at MIT new AI did lead to an attempt for humanoid AI in the Cog Project. === From Nouvelle AI to behavior-based and situated AI === The conceptual shift introduced by nouvelle AI flourished in the robotics area, given way to behavior-based robotics (BBR), a methodology for developing AI based on a modular decomposition of intelligence. It was made famous by Rodney Brooks: his subsumption architecture was one of the earliest attempts to describe a mechanism for developing BBAI. It is extremely popular in robotics and to a lesser extent to implement intelligent virtual agents because it allows the successful creation of real-time dynamic systems that can run in complex environments. For example, it underlies the intelligence of the Sony Aibo and many RoboCup robot teams. Realizing that in fact all these approaches were aiming at building not an abstract intelligence, but rather an intelligence situated in a given environment, they have come to be known as the situated approach. In fact, this approach stems out from early insights of Alan Turing, describing the need to build machines equipped with sense organs to learn directly from the real-world instead of focusing on abstract activities, such as playing chess. == Definitions == Classically, a software entity is defined as a simulated element, able to act on itself and on its environment, and which has an internal representation of itself and of the outside world. An entity can communicate with other entities, and its behavior is the consequence of its perceptions, its representations, and its interactions with the other entities. === AI loop === Simulating entities in a virtual environment requires simulating the entire process that goes from a perception of the environment, or more generally from a stimulus, to an action on the environment. This process is called the AI loop and technology used to simulate it can be subdivided in two categories. Sensorimotor or low-level AI deals with either the perception problem (what is perceived?) or the animation problem (how are actions executed?). Decisional or high-level AI deals with the action selection problem (what is the most appropriate action in response to a given perception, i.e. what is the most appropriate behavior?). === Traditional or symbolic AI === There are two main approaches in decisional AI. The vast majority of the technologies available on the market, such as planning algorithms, finite-state machines (FSA), or expert systems, are based on the traditional or symbolic AI approach. Its main characteristics are: It is top-down: it subdivides, in a recursive manner, a given problem into a series of sub-problems that are supposedly easier to solve. It is knowledge-based: it relies on a symbolic description of the world, such as a set of rules. However, the limits of traditional AI, which goal is to build systems that mimic human intelligence, are well-known: inevitably, a combinatorial explosion of the number of rules occurs due to the complexity of the environment. In fact, it is impossible to predict all the situations that will be encountered by an autonomous entity. === Situated or behavioral AI === In order to address these issues, another approach to decisional AI, also known as situated or behavioral AI, has been proposed. It does not attempt to model systems that produce deductive reasoning processes, but rather systems that behave realistically in their environment. The main characteristics of this approach are the following: It is bottom-up: it relies on elementary behaviors, which can be combined to implement more complex behaviors. It is behavior-based: it does not rely on a symbolic description of the environment, but rather on a model of the interactions of the entities with their environment. The goal of situated AI is to model entities that are autonomous in their environment. This is achieved thanks to both the intrinsic robustness of the control architecture, and its adaptation capabilities to unforeseen situations. === Situated agents === In artificial intelligence and cognitive science, the term situated refers to an agent which is embedded in an environment. The term situated is commonly used to refer to robots, but some researchers argue that software agents can also be situated if: they exist in a dynamic (rapidly changing) environment, which they can manipulate or change through their actions, and which they can sense or perceive. Examples might include web-based agents, which can alter data or trigger processes (such as purchases) over the Internet, or virtual-reality bots which inhabit and change virtual worlds, such as Second Life. Being situated is generally considered to be part of being embodied, but it is useful to consider each perspective individually. The situated perspective emphasizes that intelligent behavior derives from the environment and the agent's interactions with it. The nature of these interactions are defined by an agent's embodiment. == Implementation principles == === Modular decomposition === The most important attribute of a system driven by situated AI is that the intelligence is controlled by a set of independent semi-autonomous modules. In the original systems, each module was actually a separate device or was at least conceived of as running on its own processing thread. Generally, though, the modules are just abstractions. In this respect, situated AI may be seen as a software engineering approach to AI, perhaps akin to object oriented design. Situated AI is often associated with reactive planning, but the two are not synonymous. Brooks advocated an extreme version of cognitive minimalism which required initially that the behavior modules were finite-state machines and thus contained no conventional memory or learning. This is associated with reactive AI because reactive AI requires reacting to the current state of the world, not to an agent's memory or preconception of that world. However, learning is obviously key to realistic strong AI, so this constraint has been relaxed, though not entirely abandoned. === Action selection mechanism === The situated AI community has presented several solutions to modeling decision-making processes, also known as action selection mechanisms. The first attempt to solve this problem goes back to subsumption architectures, which were in fact more an implementation technique than an algorithm. However, this attempt paved the way to several others, in particular the free-flow hierarchies and activation networks. A comparison of the structure and performances of these two mechanisms demonstrated the advantage of using free-flow hierarchies in solving the action selection problem. However, motor schemas and process description languages are two other approaches that have been used with success for autonomous robots. == Notes and references == Arsenio, Artur M. (2004) Towards an embodied and situated AI, In: Proceedings of the International FLAIRS conference, 2004. (online) The Artificial Life Route To Artificial Intelligence: Building Embodied, Situated Agents, Luc Steels and Rodney Brooks Eds., Lawrence Erlbaum Publishing, 1995. (ISBN 978-0805815184) Rodney A. Brooks Cambrian Intelligence (MIT Press, 1999) ISBN 0-262-52263-2; collection of early papers including "Intelligence without representation" and "Intelligence without reason", from 1986 & 1991 respectively. Ronald C. Arkin Behavior-Based Robotics (MIT Press, 1998) ISBN 0-262-01165-4 Hendriks-Jansen, Horst (1996) Catching Ourselves in the Act: Situated Activity, Interactive Emergence, Evolution, and Human Thought. Cambridge, Mass.: MIT Press.

  Read more →
• 

  Region Based Convolutional Neural Networks

  Region-based Convolutional Neural Networks (R-CNN) are a family of machine learning models for computer vision, and specifically object detection and localization. The original goal of R-CNN was to take an input image and produce a set of bounding boxes as output, where each bounding box contains an object and also the category (e.g. car or pedestrian) of the object. In general, R-CNN architectures perform selective search over feature maps outputted by a CNN. R-CNN has been extended to perform other computer vision tasks, such as: tracking objects from a drone-mounted camera, locating text in an image, and enabling object detection in Google Lens. Mask R-CNN is also one of seven tasks in the MLPerf Training Benchmark, which is a competition to speed up the training of neural networks. == History == The following covers some of the versions of R-CNN that have been developed. November 2013: R-CNN. April 2015: Fast R-CNN. June 2015: Faster R-CNN. March 2017: Mask R-CNN. December 2017: Cascade R-CNN is trained with increasing Intersection over Union (IoU, also known as the Jaccard index) thresholds, making each stage more selective against nearby false positives. June 2019: Mesh R-CNN adds the ability to generate a 3D mesh from a 2D image. == Architecture == For review articles see. === Selective search === Given an image (or an image-like feature map), selective search (also called Hierarchical Grouping) first segments the image by the algorithm in (Felzenszwalb and Huttenlocher, 2004), then performs the following: Input: (colour) image Output: Set of object location hypotheses L Segment image into initial regions R = {r1, ..., rn} using Felzenszwalb and Huttenlocher (2004) Initialise similarity set S = ∅ foreach Neighbouring region pair (ri, rj) do Calculate similarity s(ri, rj) S = S ∪ s(ri, rj) while S ≠ ∅ do Get highest similarity s(ri, rj) = max(S) Merge corresponding regions rt = ri ∪ rj Remove similarities regarding ri: S = S \ s(ri, r∗) Remove similarities regarding rj: S = S \ s(r∗, rj) Calculate similarity set St between rt and its neighbours S = S ∪ St R = R ∪ rt Extract object location boxes L from all regions in R === R-CNN === With R-CNN, prediction follows a two-step process. A preprocessing selective search step generates a large set of candidate objects (typically as many as 2000), known as regions of interest (ROI). These are forwarded to a CNN, which predicts an object class score and bounding box estimate, independently for each ROI. Importantly, the ROIs are heavily filtered to remove excess candidates. This is achieved using two mechanism. Filtering begins by removing ROIs assigned to the background category. This is a specialized category, which is scored by the CNN alongside other categories. An unfortunate reality is that remaining ROIs typically suffer from heavy duplication. Namely, multiple ROIs that cover same objects in the image are all assigned non-background categories. This is resolved by a heuristic non-maximum suppression (NMS) step. === Fast R-CNN === While the original R-CNN independently computed the neural network features on each of as many as two thousand regions of interest, Fast R-CNN runs the neural network once on the whole image. At the end of the network is a ROIPooling module, which slices out each ROI from the network's output tensor, reshapes it, and classifies it. As in the original R-CNN, the Fast R-CNN uses selective search to generate its region proposals. === Faster R-CNN === While Fast R-CNN used selective search to generate ROIs, Faster R-CNN integrates the ROI generation into the neural network itself. === Mask R-CNN === While previous versions of R-CNN focused on object detections, Mask R-CNN adds instance segmentation. Mask R-CNN also replaced ROIPooling with a new method called ROIAlign, which can represent fractions of a pixel.

  Read more →
• 

  Tesla Dojo

  Tesla Dojo is a series of supercomputers designed and built by Tesla for computer vision video processing and recognition. It was used for training Tesla's machine learning models to improve its Full Self-Driving (FSD) advanced driver-assistance system. It went into production in July 2023. Dojo's goal was to efficiently process millions of terabytes of video data captured from real-life driving situations from Tesla's 4+ million cars. This goal led to a considerably different architecture than conventional supercomputer designs. In August 2025, Bloomberg News reported that the Dojo project had been disbanded, though it was restarted in January 2026. == History == Tesla operates several massively parallel computing clusters for developing its Autopilot advanced driver assistance system. Its primary unnamed cluster using 5,760 Nvidia A100 graphics processing units (GPUs) was touted by Andrej Karpathy in 2021 at the fourth International Joint Conference on Computer Vision and Pattern Recognition (CCVPR 2021) to be "roughly the number five supercomputer in the world" at approximately 81.6 petaflops, based on scaling the performance of the Nvidia Selene supercomputer, which uses similar components. However, the performance of the primary Tesla GPU cluster has been disputed, as it was not clear if this was measured using single-precision or double-precision floating point numbers (FP32 or FP64). Tesla also operates a second 4,032 GPU cluster for training and a third 1,752 GPU cluster for automatic labeling of objects. The primary unnamed Tesla GPU cluster has been used for processing one million video clips, each ten seconds long, taken from Tesla Autopilot cameras operating in Tesla cars in the real world, running at 36 frames per second. Collectively, these video clips contained six billion object labels, with depth and velocity data; the total size of the data set was 1.5 petabytes. This data set was used for training a neural network intended to help Autopilot computers in Tesla cars understand roads. By August 2022, Tesla had upgraded the primary GPU cluster to 7,360 GPUs. Dojo was first mentioned by Elon Musk in April 2019 during Tesla's "Autonomy Investor Day". In August 2020, Musk stated it was "about a year away" due to power and thermal issues. Dojo was officially announced at Tesla's Artificial Intelligence (AI) Day on August 19, 2021. Tesla revealed details of the D1 chip and its plans for "Project Dojo", a datacenter that would house 3,000 D1 chips; the first "Training Tile" had been completed and delivered the week before. In October 2021, Tesla released a "Dojo Technology" whitepaper describing the Configurable Float8 (CFloat8) and Configurable Float16 (CFloat16) floating point formats and arithmetic operations as an extension of Institute of Electrical and Electronics Engineers (IEEE) standard 754. At the follow-up AI Day in September 2022, Tesla announced it had built several System Trays and one Cabinet. During a test, the company stated that Project Dojo drew 2.3 megawatts (MW) of power before tripping a local San Jose, California power substation. At the time, Tesla was assembling one Training Tile per day. In August 2023, Tesla powered on Dojo for production use as well as a new training cluster configured with 10,000 Nvidia H100 GPUs. In January 2024, Musk described Dojo as "a long shot worth taking because the payoff is potentially very high. But it's not something that is a high probability." In June 2024, Musk explained that ongoing construction work at Gigafactory Texas is for a computing cluster claiming that it is planned to comprise an even mix of "Tesla AI" and Nvidia/other hardware with a total thermal design power of at first 130 MW and eventually exceeding 500 MW. In August 2025, Bloomberg News reported that the Dojo project was disbanded, though Musk announced it would be restarted in January 2026 with a new chip iteration. == Technical architecture == The fundamental unit of the Dojo supercomputer is the D1 chip, designed by a team at Tesla led by ex-AMD CPU designer Ganesh Venkataramanan, including Emil Talpes, Debjit Das Sarma, Douglas Williams, Bill Chang, and Rajiv Kurian. The D1 chip is manufactured by the Taiwan Semiconductor Manufacturing Company (TSMC) using 7 nanometer (nm) semiconductor nodes, has 50 billion transistors and a large die size of 645 mm2 (1.0 square inch). Updating at Artificial Intelligence (AI) Day in 2022, Tesla announced that Dojo would scale by deploying multiple ExaPODs, in which there would be: 10 Cabinets per ExaPOD (1,062,000 cores, 3,000 D1 chips) 2 System Trays per Cabinet (106,200 cores, 300 D1 chips) 6 Training Tiles per System Tray (53,100 cores, along with host interface hardware) 25 D1 chips per Training Tile (8,850 cores) 354 computing cores per D1 chip According to Venkataramanan, Tesla's senior director of Autopilot hardware, Dojo will have more than an exaflop (a million teraflops) of computing power. For comparison, according to Nvidia, in August 2021, the (pre-Dojo) Tesla AI-training center used 720 nodes, each with eight Nvidia A100 Tensor Core GPUs for 5,760 GPUs in total, providing up to 1.8 exaflops of performance. === D1 chip === Each node (computing core) of the D1 processing chip is a general purpose 64-bit CPU with a superscalar core. It supports internal instruction-level parallelism, and includes simultaneous multithreading (SMT). It doesn't support virtual memory and uses limited memory protection mechanisms. Dojo software/applications manage chip resources. The D1 instruction set supports both 64-bit scalar and 64-byte single instruction, multiple data (SIMD) vector instructions. The integer unit mixes reduced instruction set computer (RISC-V) and custom instructions, supporting 8, 16, 32, or 64 bit integers. The custom vector math unit is optimized for machine learning kernels and supports multiple data formats, with a mix of precisions and numerical ranges, many of which are compiler composable. Up to 16 vector formats can be used simultaneously. ==== Node ==== Each D1 node uses a 32-byte fetch window holding up to eight instructions. These instructions are fed to an eight-wide decoder which supports two threads per cycle, followed by a four-wide, four-way SMT scalar scheduler that has two integer units, two address units, and one register file per thread. Vector instructions are passed further down the pipeline to a dedicated vector scheduler with two-way SMT, which feeds either a 64-byte SIMD unit or four 8×8×4 matrix multiplication units. The network on-chip (NOC) router links cores into a two-dimensional mesh network. It can send one packet in and one packet out in all four directions to/from each neighbor node, along with one 64-byte read and one 64-byte write to local SRAM per clock cycle. Hardware native operations transfer data, semaphores and barrier constraints across memories and CPUs. System-wide double data rate 4 (DDR4) synchronous dynamic random-access memory (SDRAM) memory works like bulk storage. ==== Memory ==== Each core has a 1.25 megabytes (MB) of SRAM main memory. Load and store speeds reach 400 gigabytes (GB) per second and 270 GB/sec, respectively. The chip has explicit core-to-core data transfer instructions. Each SRAM has a unique list parser that feeds a pair of decoders and a gather engine that feeds the vector register file, which together can directly transfer information across nodes. ==== Die ==== Twelve nodes (cores) are grouped into a local block. Nodes are arranged in an 18×20 array on a single die, of which 354 cores are available for applications. The die runs at 2 gigahertz (GHz) and totals 440 MB of SRAM (360 cores × 1.25 MB/core). It reaches 376 teraflops using 16-bit brain floating point (BF16) numbers or using configurable 8-bit floating point (CFloat8) numbers, which is a Tesla proposal, and 22 teraflops at FP32. Each die comprises 576 bi-directional serializer/deserializer (SerDes) channels along the perimeter to link to other dies, and moves 8 TB/sec across all four die edges. Each D1 chip has a thermal design power of approximately 400 watts. === Training Tile === The water-cooled Training Tile packages 25 D1 chips into a 5×5 array. Each tile supports 36 TB/sec of aggregate bandwidth via 40 input/output (I/O) chips - half the bandwidth of the chip mesh network. Each tile supports 10 TB/sec of on-tile bandwidth. Each tile has 11 GB of SRAM memory (25 D1 chips × 360 cores/D1 × 1.25 MB/core). Each tile achieves 9 petaflops at BF16/CFloat8 precision (25 D1 chips × 376 TFLOP/D1). Each tile consumes 15 kilowatts; 288 amperes at 52 volts. === System Tray === Six tiles are aggregated into a System Tray, which is integrated with a host interface. Each host interface includes 512 x86 cores, providing a Linux-based user environment. Previously, the Dojo System Tray was known as the Training Matrix, which includes six Training Tiles, 20 Dojo Interface Processor cards across four host servers, and Ethernet-l

  Read more →
• 

  Chinchilla (language model)

  Chinchilla is a family of large language models (LLMs) developed by the research team at Google DeepMind, presented in March 2022. == Models == It is named "chinchilla" because it is a further development over a previous model family named Gopher. Both model families were trained in order to investigate the scaling laws of large language models. It claimed to outperform GPT-3. It considerably simplifies downstream utilization because it requires much less computer power for inference and fine-tuning. Based on the training of previously employed language models, it has been determined that if one doubles the model size, one must also have twice the number of training tokens. This hypothesis has been used to train Chinchilla by DeepMind. Similar to Gopher in terms of cost, Chinchilla has 70B parameters and four times as much data. Chinchilla has an average accuracy of 67.5% on the Measuring Massive Multitask Language Understanding (MMLU) benchmark, which is 7% higher than Gopher's performance. Chinchilla was still in the testing phase as of January 12, 2023. Chinchilla contributes to developing an effective training paradigm for large autoregressive language models with limited compute resources. The Chinchilla team recommends that the number of training tokens is twice for every model size doubling, meaning that using larger, higher-quality training datasets can lead to better results on downstream tasks. It has been used for the Flamingo vision-language model. == Architecture == Both the Gopher family and Chinchilla family are families of transformer models. In particular, they are essentially the same as GPT-2, with different sizes and minor modifications. Gopher family uses RMSNorm instead of LayerNorm; relative positional encoding rather than absolute positional encoding. The Chinchilla family is the same as the Gopher family, but trained with AdamW instead of Adam optimizer. The Gopher family contains six models of increasing size, from 44 million parameters to 280 billion parameters. They refer to the largest one as "Gopher" by default. Similar naming conventions apply for the Chinchilla family. Table 1 of shows the entire Gopher family: Table 4 of compares the 70-billion-parameter Chinchilla with Gopher 280B.

  Read more →
• 

  Machine learning

  Machine learning (ML) is a field of study in artificial intelligence concerned with the development and study of statistical algorithms that can learn from data and generalize to unseen data, and thus perform tasks without being explicitly programmed. Advances in the field of deep learning have allowed neural networks, a class of statistical algorithms, to surpass many previous machine learning approaches in performance. Statistics and mathematical optimisation methods compose the foundations of machine learning. Data mining is a related field of study, focusing on exploratory data analysis (EDA) through unsupervised learning. From a theoretical viewpoint, probably approximately correct learning provides a mathematical and statistical framework for describing machine learning. Most traditional machine learning and deep learning algorithms can be described as empirical risk minimisation under this framework. == History == The term machine learning was coined in 1959 by Arthur Samuel, an IBM employee and pioneer in the field of computer gaming and artificial intelligence. The synonym self-teaching computers was also used during this time period. The earliest machine learning program was introduced in the 1950s, when Samuel invented a computer program that calculated the chance of winning in checkers for each side, but the history of machine learning is rooted in decades of efforts to study human cognitive processes. In 1949, Canadian psychologist Donald Hebb published the book The Organization of Behavior, in which he introduced a theoretical neural structure formed by certain interactions among nerve cells. The Hebbian theory of neuron interaction set the groundwork for how many machine learning algorithms work, with connected artificial neurons changing the strength of their connections based on data. Other researchers who have studied human cognitive systems contributed to the modern machine learning technologies as well, including Walter Pitts and Warren McCulloch, who proposed the first mathematical model of neural networks including algorithms that mirror human thought processes. By the early 1960s, an experimental "learning machine" with punched tape memory, called Cybertron, had been developed by Raytheon Company to analyse sonar signals, electrocardiograms, and speech patterns using rudimentary reinforcement learning. It was repetitively "trained" by a human operator/teacher to recognise patterns and equipped with a "goof" button to cause it to reevaluate incorrect decisions. A representative book on research into machine learning during the 1960s was Nils Nilsson's book "Learning Machines", dealing mostly with machine learning for pattern classification. Interest related to pattern recognition continued into the 1970s, as described by Duda and Hart in 1973. In 1981, a report was given on using teaching strategies so that an artificial neural network learns to recognise 40 characters (26 letters, 10 digits, and 4 special symbols) from a computer terminal. Tom M. Mitchell provided a widely quoted, more formal definition of the algorithms studied in the machine learning field: "A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P if its performance at tasks in T, as measured by P, improves with experience E." This definition of the tasks in which machine learning is concerned is fundamentally operational rather than defining the field in cognitive terms. This follows Alan Turing's proposal in his paper "Computing Machinery and Intelligence", in which the question, "Can machines think?", is replaced by asking whether machines can convincingly imitate a human in its responses to human-posed questions. In 2014 Ian Goodfellow and others introduced generative adversarial networks (GANs) which could produce realistic synthetic data. By 2016 AlphaGo had won against top human players in Go using reinforcement learning techniques. == Relationships to other fields == === Artificial intelligence === As a scientific endeavour, machine learning grew out of the quest for artificial intelligence (AI). In the early days of AI as an academic discipline, some researchers were interested in having machines learn from data. They attempted to approach the problem with various symbolic methods, as well as what were then termed "neural networks"; these were mostly perceptrons and other models that were later found to be reinventions of the generalised linear models of statistics. Probabilistic reasoning was also employed, especially in automated medical diagnosis. However, an increasing emphasis on the logical, knowledge-based approach caused a rift between AI and machine learning. Probabilistic systems were plagued by theoretical and practical problems of data acquisition and representation. By 1980, expert systems had come to dominate AI, and statistics was out of favour. Work on symbolic/knowledge-based learning continued within AI, leading to inductive logic programming (ILP), but the more statistical line of research was now outside the field of AI proper, in pattern recognition and information retrieval. Neural network research was abandoned by AI and computer science around the same time. This subfield, termed "connectionism", was continued by researchers from other disciplines, including John Hopfield, David Rumelhart, and Geoffrey Hinton. Their main success came in the mid-1980s with the reinvention of backpropagation. Machine learning (ML), reorganised and recognised as its own field, started to flourish in the 1990s. The field changed its goal from achieving artificial intelligence to tackling solvable problems of a practical nature. It shifted focus away from the symbolic approaches it had inherited from AI, and toward methods and models borrowed from statistics, fuzzy logic, and probability theory. === Data compression === === Data mining === Machine learning and data mining often employ the same methods and overlap significantly, but while machine learning focuses on prediction based on known properties learned from the training data, data mining focuses on the discovery of previously unknown properties in the data (this is the analysis step of knowledge discovery in databases). Data mining uses many machine learning methods, but with different goals; on the other hand, machine learning also employs data mining methods as "unsupervised learning" or as a preprocessing step to improve learner accuracy. Much of the confusion between these two research communities comes from the basic assumptions they work with: in machine learning, performance is usually evaluated with respect to the ability to reproduce known knowledge, while in knowledge discovery and data mining (KDD) the key task is the discovery of previously unknown knowledge. Evaluated with respect to known knowledge, an uninformed (unsupervised) method will easily be outperformed by other supervised methods, while in a typical KDD task, supervised methods cannot be used due to the unavailability of training data. Machine learning also has intimate ties to optimization: Many learning problems are formulated as minimisation of some loss function on a training set of examples. Loss functions express the discrepancy between the predictions of the model being trained and the actual problem instances (for example, in classification, one wants to assign a label to instances, and models are trained to correctly predict the preassigned labels of a set of examples). === Generalization === Characterizing the generalisation of various learning algorithms is an active topic of current research, especially for deep learning algorithms. === Statistics === Machine learning and statistics are closely related fields in terms of methods, but distinct in their principal goal: statistics draws population inferences from a sample, while machine learning finds generalisable predictive patterns. Conventional statistical analyses require the a priori selection of a model most suitable for the study data set. In addition, only significant or theoretically relevant variables based on previous experience are included for analysis. In contrast, machine learning is not built on a pre-structured model; rather, the data shape the model by detecting underlying patterns. The more variables (input) used to train the model, the more accurate the ultimate model will be. Leo Breiman distinguished two statistical modelling paradigms: the data model and the algorithmic model, wherein "algorithmic model" means more or less the machine learning algorithms like Random forest. Some statisticians have adopted methods from machine learning, producing the field of statistical learning. === Statistical physics === Analytical and computational techniques derived from deep-rooted physics of disordered systems can be extended to large-scale problems, including machine learning, e.g., to analyse the weight space of deep neural networks. Statistical physics is thus

  Read more →
• 

  Underwater computer vision

  Underwater computer vision is a subfield of computer vision. In recent years, with the development of underwater vehicles ( ROV, AUV, gliders), the need to be able to record and process huge amounts of information has become increasingly important. Applications range from inspection of underwater structures for the offshore industry to the identification and counting of fishes for biological research. However, no matter how big the impact of this technology can be to industry and research, it still is in a very early stage of development compared to traditional computer vision. One reason for this is that, the moment the camera goes into the water, a whole new set of challenges appear. On one hand, cameras have to be made waterproof, marine corrosion deteriorates materials quickly and access and modifications to experimental setups are costly, both in time and resources. On the other hand, the physical properties of the water make light behave differently, changing the appearance of a same object with variations of depth, organic material, currents, temperature etc. == Applications == Seafloor survey Vehicle navigation and positioning Biological monitoring {possibly aquatic biomonitoring) Video mosaics as visual navigation maps Submarine pipeline inspection Wreckage visualization Maintenance of underwater structures Drowning detection systems == Medium differences == === Illumination === In air, light comes from the whole hemisphere on cloudy days, and is dominated by the sun. In water direct lighting comes from a cone about 96° wide above the scene. This phenomenon is called Snell's window. Artificial lighting can be used where natural light levels are insufficient and where the light path is too long to produce acceptable colour, as the loss of colour is a function of the total distance through water from the source to the camera lens port. === Light attenuation === Unlike air, water attenuates light exponentially. This results in hazy images with very low contrast. The main reasons for light attenuation are light absorption (where energy is removed from the light) and light scattering, by which the direction of light is changed. Light scattering can further be divided into forward scattering, which results in an increased blurriness and backward scattering that limits the contrast and is responsible for the characteristic veil of underwater images. Both scattering and attenuation are heavily influenced by the amount of organic matter dissolved or suspended in the water. Light attenuation in water is also a function of the wavelength. This means that different colours are attenuated at different rates, leading to colour degradation.with depth and distance. Red and orange light are attenuated faster, followed by yellows and greens. Blue is the least attenuated visible wavelength. === Artificial lighting === == Challenges == In high level computer vision, human structures are frequently used as image features for image matching in different applications. However, the sea bottom lacks such features, making it hard to find correspondences in two images. In order to be able to use a camera in the water, a watertight housing is required. However, refraction will happen at the water-glass and glass-air interface due to differences in density of the materials. This has the effect of introducing a non-linear image deformation. The motion of the vehicle presents another special challenge. Underwater vehicles are constantly moving due to currents and other phenomena. This introduces another uncertainty to algorithms, where small motions may appear in all directions. This can be specially important for video tracking. In order to reduce this problem image stabilization algorithms may be applied. == Relevant technology == === Image restoration === Image restoration< techniques are intended to model the degradation process and then invert it, obtaining the new image after solving. It is generally a complex approach that requires plenty of parameters that vary a lot between different water conditions. === Image enhancement === Image enhancement only tries to provide a visually more appealing image without taking the physical image formation process into account. These methods are usually simpler and less computational intensive. === Color correction === Various algorithms exist that perform automatic color correction. The UCM (Unsupervised Color Correction Method), for example, does this in the following steps: It firstly reduces the color cast by equalizing the color values. Then it enhances contrast by stretching the red histogram towards the maximum and finally saturation and intensity components are optimized. == Underwater stereo vision == It is usually assumed that stereo cameras have been calibrated previously, geometrically and radiometrically. This leads to the assumption that corresponding pixels should have the same color. However this can not be guaranteed in an underwater scene, because of dispersion and backscatter. However, it is possible to digitally model this phenomenon and create a virtual image with those effects removed == Other application fields == Imaging sonars have become more and more accessible and gained resolution, delivering better images. Sidescan sonars are used to produce complete maps of regions of the sea floor stitching together sequences of sonar images. However, sonar images often lack proper contrast and are degraded by artefacts and distortions due to noise, attitude changes of the AUV/ROV carrying the sonar or non uniform beam patterns. Another common problem with sonar computer vision is the comparatively low frame rate of sonar images.

  Read more →
• 

  Speech segmentation

  Speech segmentation is the process of identifying the boundaries between words, syllables, or phonemes in spoken natural languages. The term applies both to the mental processes used by humans, and to artificial processes of natural language processing. In the field of automatic pronunciation assessment, the process of segmenting an utterance against expected word(s) is called forced alignment. Speech segmentation is a subfield of general speech perception and an important subproblem of the technologically focused field of speech recognition, and cannot be adequately solved in isolation. As in most natural language processing problems, one must take into account context, grammar, and semantics, and even so the result is often a probabilistic division (statistically based on likelihood) rather than a categorical one. Though it seems that coarticulation—a phenomenon which may happen between adjacent words just as easily as within a single word—presents the main challenge in speech segmentation across languages, some other problems and strategies employed in solving those problems can be seen in the following sections. This problem overlaps to some extent with the problem of text segmentation that occurs in some languages which are traditionally written without inter-word spaces, like Chinese and Japanese, compared to writing systems which indicate speech segmentation between words by a word divider, such as the space. However, even for those languages, text segmentation is often much easier than speech segmentation, because the written language usually has little interference between adjacent words, and often contains additional clues not present in speech (such as the use of Chinese characters for word stems in Japanese). == Lexical recognition == In natural languages, the meaning of a complex spoken sentence can be understood by decomposing it into smaller lexical segments (roughly, the words of the language), associating a meaning to each segment, and combining those meanings according to the grammar rules of the language. Though lexical recognition is not thought to be used by infants in their first year, due to their highly limited vocabularies, it is one of the major processes involved in speech segmentation for adults. Three main models of lexical recognition exist in current research: first, whole-word access, which argues that words have a whole-word representation in the lexicon; second, decomposition, which argues that morphologically complex words are broken down into their morphemes (roots, stems, inflections, etc.) and then interpreted and; third, the view that whole-word and decomposition models are both used, but that the whole-word model provides some computational advantages and is therefore dominant in lexical recognition. To give an example, in a whole-word model, the word "cats" might be stored and searched for by letter, first "c", then "ca", "cat", and finally "cats". The same word, in a decompositional model, would likely be stored under the root word "cat" and could be searched for after removing the "s" suffix. "Falling", similarly, would be stored as "fall" and suffixed with the "ing" inflection. Though proponents of the decompositional model recognize that a morpheme-by-morpheme analysis may require significantly more computation, they argue that the unpacking of morphological information is necessary for other processes (such as syntactic structure) which may occur parallel to lexical searches. As a whole, research into systems of human lexical recognition is limited due to little experimental evidence that fully discriminates between the three main models. In any case, lexical recognition likely contributes significantly to speech segmentation through the contextual clues it provides, given that it is a heavily probabilistic system—based on the statistical likelihood of certain words or constituents occurring together. For example, one can imagine a situation where a person might say "I bought my dog at a ____ shop" and the missing word's vowel is pronounced as in "net", "sweat", or "pet". While the probability of "netshop" is extremely low, since "netshop" isn't currently a compound or phrase in English, and "sweatshop" also seems contextually improbable, "pet shop" is a good fit because it is a common phrase and is also related to the word "dog". Moreover, an utterance can have different meanings depending on how it is split into words. A popular example, often quoted in the field, is the phrase "How to wreck a nice beach", which sounds very similar to "How to recognize speech". As this example shows, proper lexical segmentation depends on context and semantics which draws on the whole of human knowledge and experience, and would thus require advanced pattern recognition and artificial intelligence technologies to be implemented on a computer. Lexical recognition is of particular value in the field of computer speech recognition, since the ability to build and search a network of semantically connected ideas would greatly increase the effectiveness of speech-recognition software. Statistical models can be used to segment and align recorded speech to words or phones. Applications include automatic lip-synch timing for cartoon animation, follow-the-bouncing-ball video sub-titling, and linguistic research. Automatic segmentation and alignment software is commercially available. == Phonotactic cues == For most spoken languages, the boundaries between lexical units are difficult to identify; phonotactics are one answer to this issue. One might expect that the inter-word spaces used by many written languages like English or Spanish would correspond to pauses in their spoken version, but that is true only in very slow speech, when the speaker deliberately inserts those pauses. In normal speech, one typically finds many consecutive words being said with no pauses between them, and often the final sounds of one word blend smoothly or fuse with the initial sounds of the next word. The notion that speech is produced like writing, as a sequence of distinct vowels and consonants, may be a relic of alphabetic heritage for some language communities. In fact, the way vowels are produced depends on the surrounding consonants just as consonants are affected by surrounding vowels; this is called coarticulation. For example, in the word "kit", the [k] is farther forward than when we say 'caught'. But also, the vowel in "kick" is phonetically different from the vowel in "kit", though we normally do not hear this. In addition, there are language-specific changes which occur in casual speech which makes it quite different from spelling. For example, in English, the phrase "hit you" could often be more appropriately spelled "hitcha". From a decompositional perspective, in many cases, phonotactics play a part in letting speakers know where to draw word boundaries. In English, the word "strawberry" is perceived by speakers as consisting (phonetically) of two parts: "straw" and "berry". Other interpretations such as "stra" and "wberry" are inhibited by English phonotactics, which does not allow the cluster "wb" word-initially. Other such examples are "day/dream" and "mile/stone" which are unlikely to be interpreted as "da/ydream" or "mil/estone" due to the phonotactic probability or improbability of certain clusters. The sentence "Five women left", which could be phonetically transcribed as [faɪvwɪmɘnlɛft], is marked since neither /vw/ in /faɪvwɪmɘn/ nor /nl/ in /wɪmɘnlɛft/ are allowed as syllable onsets or codas in English phonotactics. These phonotactic cues often allow speakers to easily distinguish the boundaries in words. Vowel harmony in languages like Finnish can also serve to provide phonotactic cues. While the system does not allow front vowels and back vowels to exist together within one morpheme, compounds allow two morphemes to maintain their own vowel harmony while coexisting in a word. Therefore, in compounds such as "selkä/ongelma" ('back problem') where vowel harmony is distinct between two constituents in a compound, the boundary will be wherever the switch in harmony takes place—between the "ä" and the "ö" in this case. Still, there are instances where phonotactics may not aid in segmentation. Words with unclear clusters or uncontrasted vowel harmony as in "opinto/uudistus" ('student reform') do not offer phonotactic clues as to how they are segmented. From the perspective of the whole-word model, however, these words are thought be stored as full words, so the constituent parts would not necessarily be relevant to lexical recognition. == In infants and non-natives == Infants are one major focus of research in speech segmentation. Since infants have not yet acquired a lexicon capable of providing extensive contextual clues or probability-based word searches within their first year, as mentioned above, they must often rely primarily upon phonotactic and rhythmic cues (with prosody being the dominant cue), all

  Read more →
• 

  NetOwl

  NetOwl is a suite of multilingual text and identity analytics products that analyze big data in the form of text data – reports, web, social media, etc. – as well as structured entity data about people, organizations, places, and things. NetOwl utilizes artificial intelligence (AI)-based approaches, including natural language processing (NLP), machine learning (ML), and computational linguistics, to extract entities, relationships, and events; to perform sentiment analysis; to assign latitude/longitude to geographical references in text; to translate names written in foreign languages; and to perform name matching and identity resolution. NetOwl's uses include semantic search and discovery, geospatial analysis, intelligence analysis, content enrichment, compliance monitoring, cyber threat monitoring, risk management, and bioinformatics. == History == The first NetOwl product was NetOwl Extractor, which was initially released in 1996. Since then, Extractor has added many new capabilities, including relationship and event extraction, categorization, name translation, geotagging, and sentiment analysis, as well as entity extraction in other languages. Other products were added later to the NetOwl suite, namely TextMiner, NameMatcher, and EntityMatcher. NetOwl has participated in several 3rd party-sponsored text and entity analytics software benchmarking events. NetOwl Extractor was the top-scoring named entity extraction system at the DARPA-sponsored Message Understanding Conference MUC-6 and the top-scoring link and event extraction system in MUC-7. It was also the top-scoring system at several of the NIST-sponsored Automatic Content Extraction (ACE) evaluation tasks. NetOwl NameMatcher was the top-scoring system at the MITRE Challenge for Multicultural Person Name Matching. == Products == The NetOwl suite includes, among others, the following text and entity analytics products: === Text Analytics === NetOwl Extractor performs entity extraction from unstructured texts using natural language processing (NLP), machine learning (ML), and computational linguistics. Extractor also performs semantic relationship and event extraction as well as geotagging of text. It is used for a variety of data sources including both traditional sources (e.g., news, reports, web pages, email) and social media (e.g., Twitter, Facebook, chats, blogs). It runs on a variety of Big Data analytics platforms, including Apache Hadoop and LexisNexis’s High-Performance Computer Cluster (HPCC) technology. It has been integrated with a number of 3rd party analytical tools such as Esri ArcGIS and Google Earth/Maps. === Identity Analytics === NetOwl NameMatcher and EntityMatcher perform name matching and identity resolution for large multicultural and multilingual entity databases using machine learning (ML) and computational linguistics approaches. They are used for applications such as anti–money laundering (AML), watch lists, regulatory compliance, fraud detection, etc.

  Read more →
• 

  Open-source robotics

  Open-source robotics is a branch of robotics where robots are developed with open-source hardware and free and open-source software, publicly sharing blueprints, schematics, and source code. It is thus closely related to the open design movement, the maker movement and open science. == Requirements == Open source robotics means that information about the hardware is easily discerned, so that others can easily rebuild it. In turn, this requires design to use only easily available standard subcomponents and tools, and for the build process to be documented in detail including a bill of materials and detailed ('Ikea style') step-by-step building and testing instructions. (A CAD file alone is not sufficient, as it does not show the steps for performing or testing the build). These requirements are standard to open source hardware in general, and are formalised by various licences, certifications, especially those defined by the peer-reviewed journals Journal of Open Hardware and HardwareX. Licensing requirements for software are the same as for any open source software. But in addition, for software components to be of practical use in real robot systems, they need to be compatible with other software, usually as defined by some robotics middleware community standard. == Hardware systems == Applications to date include: Robot arms, e.g. PARA or Thor Wheeled mobile robots. e.g. OpenScout Four-legged robots such as the Open Dynamic Robot Initiative UAV quadcopters (drones) such as Agilicious Humanoid robots, e.g. iCub, Berkeley Humanoid Lite Self-driving cars, e.g. OpenPodcar (→ Personal rapid transit) Submersible robots, eg. OpenFish Laboratory robotics such as chemical liquid handling Vertical farming Swarm robots, e.g. HeRoSwarm Domestic tasks: vacuum cleaning, floor washing and grass mowing Robot sports including robot combat and autonomous racing Education == Hardware subcomponents == Most open source hardware definitions allow non-open subcomponents to be used in modular design, as long as they are easily available. However many designs try to push openness down into as many subcomponents as possible, with the aim of ultimately reaching fully open designs. Open hardware manual-drive vehicles and their subcomponents, such as from Open Source Ecology, are often used as starting points and extended with automation systems. Open subcomponents can include open-source computing hardware as subcomponents, such as Arduino and RISC-V, as well as open source motors and drivers such as the Open Source Motor Controller and ODrive. Open hardware robotics interface boards can simplify interfacing between middleware software and physical hardware. == Software subcomponents == === Middleware === Robotics middleware is software which links multiple other software components together. In robotics, this specifically means real-time communication systems with standardized message passing protocols. The predominant open source middleware is ROS2, the robot operating system, now as version 2. Other alternatives include ROS1, YARP — used in the iCub, URBI, and Orca. Open source middleware is usually run on an open source operating system, especially the Ubuntu distribution of Linux. === Driver software === Most robot sensors and actuators require software drivers. There is little standardization of open source software at this level, because each hardware device is different. Creating open drivers for closed hardware is difficult as it requires both low level programming and reverse engineering. === Simulation software === Open source robotics simulators include Gazebo, MuJoCo and Webots. Open source 3D game engines such as Godot are also sometimes used as simulators, when equipped with suitable middleware interfaces. === Automation software === At the level of AI, many standard algorithms have open source software implementations, mostly in ROS2. Major components include: Machine vision systems such as the YOLO object detector. 3D photogrammetry Navigation including SLAM and planning such as nav2 Arm inverse kinematics such as moveIt2 == Community == The first signs of the increasing popularity of building and sharing robot designs were found with the maker culture community. What began with small competitions for remote operated vehicles (e.g. Robot combat), soon developed to the building of autonomous telepresence robots such as Sparky and then true robots (being able to take decisions themselves) as the Open Automaton Project. Several commercial companies now also produce kits for making simple robots. The community has adopted open source hardware licenses, certifications, and peer-reviewed publications, which check that source has been made correctly and permanently available under community definitions, and which validate that this has been done. These processes have become critically important due to many historical projects claiming to be open source but them reverting on the promise due to commercialisation or other pressures. As with other forms of open source hardware, the community continues to debate precise criteria for 'ease of build'. A common standard is that designs should be buildable by a technical university student, in a few days, using typical fablab tools, but definitions of all of these subterms can also be debated. Compared to other forms of open source hardware, open source robotics typically includes a large software element, so involves software as well as hardware engineers. Open source concepts are more established in open source software than hardware, so robotics is a field in which those concepts can be shared and transferred from software to hardware. While the community in open source robotics is multi-faceted with a wide range of backgrounds, a sizable sub-community uses the ROS middleware and meets at the ROSCon conferences to discuss development of ROS itself and automation components built on it.

  Read more →
• 

  Euratlas

  Euratlas is a Switzerland-based software company dedicated to elaborate digital history maps of Europe. Founded in 2001, Euratlas has created a collection of history maps of Europe from year 1 AD to year 2000 AD that present the evolution of every country from the Roman Empire to present times. The evolution includes sovereign states and their administrative subdivisions, but also unorganized peoples and dependent territories. The maps show European country borders at regular intervals of 100 years, but not year by year. This leaves out many important turning points in history. Euratlas is considered a digital humanities company, and a scholar research software used in the field of historic cartography. It is broadly known among American and European universities, who mainly use Euratlas as a research tool and as a digital library atlas. == Sequential mapping policy == This concept was first designed by the German scholar Christian Kruse (1753–1827). Kruse, well aware that historical accounts are often biased for geographical, philosophical or political reasons, created a set of sequential maps in order to give a global vision of the successive political situations. Nowadays, the majority of atlases don't use this approach, but are event-based, like the well-known Penguin Atlas of History. The sequential approach intends to make the sequence of maps more neutral and suitable for students, historians and professionals of several fields. Although, this approach has been discussed as it leaves out many important history events that are not reflected on any of the maps because of the century interval. == Geo-referenced historical data == Initially, the European maps by century were developed as vector maps. From 2006 on, they have been converted to a geographic information system (GIS) database, enabling geo-referenced data capabilities. The map information is distributed in several layers: physical (geography information layer); political information layer (supranational entities, sovereign states, administrative divisions, dependent states and autonomous peoples); and special layers for cities and uncertain borders. The software database also contains much non-geographical information about political relationships between the various kinds of territories. == Map projection == Euratlas History Maps uses a Mercator projection, with the center in Europe. The maps include the North-African coast and the Near-East, offering a complete view of the Mediterranean Basin. The European Russia plains are shown, but not Scandinavia, specially Finland, which is cropped off the map view.

  Read more →
• 

  Microapp

  A microapp is a super-specialized application designed to perform one task or use case with the only objective of doing it well. They follow the single responsibility principle, which states that "a class should have one and only one reason to change." Micro applications help developers create less complex applications while reducing costs by breaking down monolithic systems into groups of independent services acting as one system. A good example of Microapps would be https://docs.citrix.com/en-us/legacy-archive/downloads/microapps.pdfthat provide single purpose action from Salesforce and over 40 applications on its workspace. == Requirements and characteristics == Microapps usually are accessible on any device, display, or operating system without installation on the viewer's device. To qualify as a microapp, the entity must: be built and deployed as an independent software module bring together various media types into a single experience have advanced security and compliance features be functionally-extensible comply with granular data demands be agnostic single use case oriented Microapps differentiate from traditional web or mobile applications by how the end-user interacts with them. Consequently, they can be embedded in websites or viewed online to bypass app stores and are typically built to provide a focused experience to the user. == Usage == Microapps are typically used for commercial purposes to reduce development costs for projects not requiring the large scope of a traditional web or mobile application. In addition, they are often used to showcase in-depth information or enrich marketing material with interactivity. Lately, micro apps are being used to boost productivity by providing quick tools to people to reuse best practices. Users have been interacting with microapps for a while with suites like Microsoft 365 and Google Workspace, where each one of their end-user services could be considered as a microapp. All these microapps share a unique identity manager to provide a unified user experience. == Benefits == Replacing monolith systems with microapps provide several advantages like: Reduce complexity for developers and users. Smaller, more cohesive, and maintainable codebases Scalable organizations with decoupled, autonomous teams Allows for hyper-specialization Independent deployment Multi-stack == Cloud-native microapps == Technologies like Kubernetes, or OpenShift, allow companies to replace their monolith and legacy systems with modular software taking advantage of microapps on reducing costs and improve reliability and security. == Microapps vs. microservices == There is a widespread misunderstanding between these two concepts, which is the key difference. Microservices is an architectural style that is systems-centric, meaning it decouples the presentation and data layer using web services APIs. On the other side, micro apps behave more as a super-architecture style (that embraces microservices among other types), and it is user-centric, meaning they decouple the whole monolith system onto modules that are designed to interact with final users. Both architectural styles rely on modularity to provide high performance, scalability, and resilience. == Considerations == Developing Micro apps requires a different approach than traditional software, and user experience is crucial. The following considerations are essential for switching to microapps. To run multiple microapps is required a single identity management system. Microservices are well suited to make microapps more powerful Apps with different levels of maturity might create a non-unified user experience. Duplication of dependencies can create security issues and inefficiencies. Suitable for well-organized teams

  Read more →
• 

  Core FTP

  Core FTP LE is a freeware secure FTP client for Windows, developed by CoreFTP.com. Features include FTP, SSL/TLS, SFTP via SSH, and HTTP/HTTPS support. Secure FTP clients encrypt account information and data transferred across the internet, protecting data from being seen, or sniffed across networks. Core FTP is a traditional FTP client with local files displayed on the left, remote files on the right. Core FTP Server is a secure FTP server for Windows, developed by CoreFTP.com, starting in 2010. == Licensing == CoreFTP LE is free for personal, educational, non-profit, and business use.

  Read more →
• 

  Model compression

  Model compression is a machine learning technique for reducing the size of trained models. Large models can achieve high accuracy, but often at the cost of significant resource requirements. Compression techniques aim to compress models without significant performance reduction. Smaller models require less storage space, and consume less memory and compute during inference. Compressed models enable deployment on resource-constrained devices such as smartphones, embedded systems, edge computing devices, and consumer electronics computers. Efficient inference is also valuable for large corporations that serve large model inference over an API, allowing them to reduce computational costs and improve response times for users. Model compression is not to be confused with knowledge distillation, in which a smaller "student" model is trained to imitate the input-output behavior of a larger "teacher" model (as opposed to using the "teacher"'s trained parameters or the "teacher"'s training targets). == Techniques == Several techniques are employed for model compression. === Pruning === Pruning sparsifies a large model by setting some parameters to exactly zero. This effectively reduces the number of parameters. This allows the use of sparse matrix operations, which are faster than dense matrix operations. Pruning criteria can be based on magnitudes of parameters, the statistical pattern of neural activations, Hessian values, etc. === Quantization === Quantization reduces the numerical precision of weights and activations. For example, instead of storing weights as 32-bit floating-point numbers, they can be represented using 8-bit integers. Low-precision parameters take up less space, and takes less compute to perform arithmetic with. It is also possible to quantize some parameters more aggressively than others, so for example, a less important parameter can have 8-bit precision while another, more important parameter, can have 16-bit precision. Inference with such models requires mixed-precision arithmetic. Quantized models can also be used during training (rather than after training). PyTorch implements automatic mixed-precision (AMP), which performs autocasting, gradient scaling, and loss scaling. === Low-rank factorization === Weight matrices can be approximated by low-rank matrices. Let W {\displaystyle W} be a weight matrix of shape m × n {\displaystyle m\times n} . A low-rank approximation is W ≈ U V T {\displaystyle W\approx UV^{T}} , where U {\displaystyle U} and V {\displaystyle V} are matrices of shapes m × k , n × k {\displaystyle m\times k,n\times k} . When k {\displaystyle k} is small, this both reduces the number of parameters needed to represent W {\displaystyle W} approximately, and accelerates matrix multiplication by W {\displaystyle W} . Low-rank approximations can be found by singular value decomposition (SVD). The choice of rank for each weight matrix is a hyperparameter, and jointly optimized as a mixed discrete-continuous optimization problem. The rank of weight matrices may also be pruned after training, taking into account the effect of activation functions like ReLU on the implicit rank of the weight matrices. == Training == Model compression may be decoupled from training, that is, a model is first trained without regard for how it might be compressed, then it is compressed. However, it may also be combined with training. The "train big, then compress" method trains a large model for a small number of training steps (less than it would be if it were trained to convergence), then heavily compress the model. It is found that at the same compute budget, this method results in a better model than lightly compressed, small models. In Deep Compression, the compression has three steps. First loop (pruning): prune all weights lower than a threshold, then finetune the network, then prune again, etc. Second loop (quantization): cluster weights, then enforce weight sharing among all weights in each cluster, then finetune the network, then cluster again, etc. Third step: Use Huffman coding to losslessly compress the model. The SqueezeNet paper reported that Deep Compression achieved a compression ratio of 35 on AlexNet, and a ratio of ~10 on SqueezeNets.

  Read more →
• 

  Document mosaicing

  Document mosaicing is a process that stitches multiple, overlapping snapshot images of a document together to produce one large, high resolution composite. The document is slid under a stationary, over-the-desk camera by hand until all parts of the document are snapshotted by the camera's field of view. As the document slid under the camera, all motion of the document is coarsely tracked by the vision system. The document is periodically snapshotted such that the successive snapshots are overlap by about 50%. The system then finds the overlapped pairs and stitches them together repeatedly until all pairs are stitched together as one piece of document. The document mosaicing can be divided into four main processes. Tracking Feature detecting Correspondences establishing Images mosaicing. == Tracking (simple correlation process) == In this process, the motion of the document slid under the camera is coarsely tracked by the system. Tracking is performed by a process called simple correlation process. In the first frame of snapshots, a small patch is extracted from the center of the image as a correlation template. The correlation process is performed in the four times size of the patch area of the next frame. The motion of the paper is indicated by the peak in the correlation function. The peak in the correlation function indicates the motion of the paper. The template is resampled from this frame and the tracking continues until the template reaches the edge of the document. After the template reaches the edge of the document, another snapshot is taken and the tracking process performs repeatedly until the whole document is imaged. The snapshots are stored in an ordered list to facilitate pairing the overlapped images in later processes. == Feature detecting for efficient matching == Feature detection is the process of finding the transformation that aligns one image with another. There are two main approaches for feature detection. Feature-based approach : Motion parameters are estimated from point correspondences. This approach is suitable for the case that there is plenty supply of stable and detectable features. Featureless approach : When the motion between the two images is small, the motion parameters are estimated using optical flow. On the other hand, when the motion between the two images is large, the motion parameters are estimated using generalised cross-correlation. However, this approach requires a computationally expensive resources. Each image is segmented into a hierarchy of columns, lines, and words to match the organised sets of features across images. Skew angle estimation and columns, lines and words finding are the examples of feature detection operations. === Skew angle estimation === Firstly, the angle that the rows of text make with the image raster lines (skew angle) is estimated. It is assumed to lie in the range of ±20°. A small patch of text in the image is selected randomly and then rotated in the range of ±20° until the variance of the pixel intensities of the patch summed along the raster lines is maximised. To ensure that the found skew angle is accurate, the document mosaic system performs calculation at many image patches and derive the final estimation by finding the average of the individual angles weighted by the variance of the pixel intensities of each patch. === Columns, lines and words finding === In this operation, the de-skewed document is intuitively segmented into a hierarchy of columns, lines and words. The sensitivity to illumination and page coloration of the de-skewed document can be removed by applying a Sobel operator to the de-skewed image and thresholding the output to obtain the binary gradient, de-skewed image. The operation can be roughly separated into 3 steps: column segmentation, line segmentation and word segmentation. Columns are easily segmented from the binary gradient, de-skewed images by summing pixels vertically. Baselines of each row are segmented in the same way as the column segmentation process but horizontally. Finally, individual words are segmented by applying the vertical process at each segmented row. These segmentations are important because the document mosaic is created by matching the lower right corners of words in overlapping images pair. Moreover, the segmentation operation can organize the list of images in the context of a hierarchy of rows and column reliably. The segmentation operation involves a considerable amount of summing in the binary gradient, de-skewed images, which done by construct a matrix of partial sums whose elements are given by p i y = ∑ u = 1 i ∑ v = 1 j b u v {\displaystyle p_{iy}=\sum _{u=1}^{i}\sum _{v=1}^{j}b_{uv}} The matrix of partial sums is calculated in one pass through the binary gradient, de-skewed image. ∑ u = u 1 u 2 ∑ v = v 1 v 2 b u v = p u 2 v 2 + p u 1 v 1 − p u 1 v 2 − p u 2 v 1 {\displaystyle \sum _{u=u_{1}}^{u_{2}}\sum _{v=v_{1}}^{v_{2}}b_{uv}=p_{u_{2}v_{2}}+p_{u_{1}v_{1}}-p_{u_{1}v_{2}}-p_{u_{2}v_{1}}} == Correspondences establishing == The two images are now organized in hierarchy of linked lists in following structure : image=list of columns row=list of words column=list of row word=length (in pixels) At the bottom of the structure, the length of each word is recorded for establishing correspondence between two images to reduce to search only the corresponding structures for the groups of words with the matching lengths. === Seed match finding === A seed match finding is done by comparing each row in image1 with each row in image2. The two rows are then compared to each other by every word. If the length (in pixel) of the two words (one from image1 and one from image2) and their immediate neighbours agree with each other within a predefined tolerance threshold (5 pixels, for example), then they are assumed to match. The row of each image is assumed a match if there are three or more word matches between the two rows. The seed match finding operation is terminated when two pairs of consecutive row match are found. === Match list building === After finishing a seed match finding operation, the next process is to build the match list to generate the correspondences points of the two images. The process is done by searching the matching pairs of rows away from the seed row. == Images mosaicing == Given the list of corresponding points of the two images, finding the transformation of the overlapping portion of the images is the next process. Assuming a pinhole camera model, the transformation between pixels (u,v) of image 1 and pixels (u0, v0) of image 2 is demonstrated by a plane-to-plane projectivity. [ s u ′ s v ′ s ] = [ p 11 p 12 p 13 p 21 p 22 p 23 p 31 p 32 1 ] [ u v 1 ] E q .1 {\displaystyle \left[{\begin{array}{c}su'\\sv'\\s\end{array}}\right]=\left[{\begin{array}{ccc}p_{11}&p_{12}&p_{13}\\p_{21}&p_{22}&p_{23}\\p_{31}&p_{32}&1\end{array}}\right]\left[{\begin{array}{c}u\\v\\1\end{array}}\right]\qquad Eq.1} The parameters of the projectivity is found from four pairs of matching points. RANSAC regression technique is used to reject outlying matches and estimate the projectivity from the remaining good matches. The projectivity is fine-tuned using correlation at the corners of the overlapping portion to obtain four correspondences to sub-pixel accuracy. Therefore, image1 is then transformed into image2's coordinate system using Eq.1. The typical result of the process is shown in Figure 5. === Many images coping === Finally, the whole page composition is built up by mapping all the images into the coordinate system of an "anchor" image, which is normally the one nearest the page center. The transformations to the anchor frame are calculated by concatenating the pair-wise transformations found earlier. The raw document mosaic is shown in Figure 6. However, there might be a problem of non-consecutive images that are overlap. This problem can be solved by performing Hierarchical sub-mosaics. As shown in Figure 7, image1 and image2 are registered, as are image3 and image4, creating two sub-mosaics. These two sub-mosaics are later stitched together in another mosaicing process. == Applied areas == There are various areas that the technique of document mosaicing can be applied to such as : Text segmentation of images of documents Document Recognition Interaction with paper on the digital desk Video mosaics for virtual environments Image registration techniques == Relevant research papers == Huang, T.S.; Netravali, A.N. (1994). "Motion and structure from feature correspondences: A review". Proceedings of the IEEE. 82 (2): 252–268. doi:10.1109/5.265351. D.G. Lowe. [1] Perceptual Organization and Visual Recognition. Kluwer Academic Publishers, Boston, 1985. Irani, M.; Peleg, S. (1991). "Improving resolution by image registration". CVGIP: Graphical Models and Image Processing. 53 (3): 231–239. doi:10.1016/1049-9652(91)90045-L. S2CID 4834546. Shivakumara, P.; Kumar, G. Hemantha; Guru, D. S.; Nagabhushan, P. (2006). "

  Read more →
• 

  SimSimi

  SimSimi is an artificial intelligence conversation program created in 2002 by ISMaker. It grows its artificial intelligence day by day assisted by a feature that allows users to teach it to respond correctly. SimSimi, pronounced as "shim-shimi", is from a Korean word simsim (심심) which means "bored". It has an application designed for Android, Windows Phone and iOS. The application was banned in Thailand in 2012 after users taught it to make responses containing profanity, and to criticise leading politicians. In April 2018, SimSimi was suspended in Brazil due to accusations of sending inappropriate messages, such as sexual language, bullying and even death threats, being labeled as "dangerous" mainly due to its popularity among children, and according to its developer, the suspension of the app in the country "was inevitable because the SimSimi app, at least in the last few days, had a significant negative social impact in Brazil.”

  Read more →