In machine learning and statistics, the learning rate is a tuning parameter in an optimization algorithm that determines the step size at each iteration while moving toward a minimum of a loss function. Since it influences to what extent newly acquired information overrides old information, it metaphorically represents the speed at which a machine learning model "learns". In the adaptive control literature, the learning rate is commonly referred to as gain. In setting a learning rate, there is a trade-off between the rate of convergence and overshooting. While the descent direction is usually determined from the gradient of the loss function, the learning rate determines how big a step is taken in that direction. Too high a learning rate will make the learning jump over minima, but too low a learning rate will either take too long to converge or get stuck in an undesirable local minimum. In order to achieve faster convergence, prevent oscillations and getting stuck in undesirable local minima the learning rate is often varied during training either in accordance to a learning rate schedule or by using an adaptive learning rate. The learning rate and its adjustments may also differ per parameter, in which case it is a diagonal matrix that can be interpreted as an approximation to the inverse of the Hessian matrix in Newton's method. The learning rate is related to the step length determined by inexact line search in quasi-Newton methods and related optimization algorithms. == Learning rate schedule == Initial rate can be left as system default or can be selected using a range of techniques. A learning rate schedule changes the learning rate during learning and is most often changed between epochs/iterations. This is mainly done with two parameters: decay and momentum. There are many different learning rate schedules but the most common are time-based, step-based and exponential. Decay serves to settle the learning in a nice place and avoid oscillations, a situation that may arise when too high a constant learning rate makes the learning jump back and forth over a minimum, and is controlled by a hyperparameter. Momentum is analogous to a ball rolling down a hill; we want the ball to settle at the lowest point of the hill (corresponding to the lowest error). Momentum both speeds up the learning (increasing the learning rate) when the error cost gradient is heading in the same direction for a long time and also avoids local minima by 'rolling over' small bumps. Momentum is controlled by a hyperparameter analogous to a ball's mass which must be chosen manually—too high and the ball will roll over minima which we wish to find, too low and it will not fulfil its purpose. The formula for factoring in the momentum is more complex than for decay but is most often built in with deep learning libraries such as Keras. Time-based learning schedules alter the learning rate depending on the learning rate of the previous time iteration. Factoring in the decay the mathematical formula for the learning rate is: η n + 1 = η 0 1 + d n {\displaystyle \eta _{n+1}={\frac {\eta _{0}}{1+dn}}} where η {\displaystyle \eta } is the learning rate, η 0 {\displaystyle \eta _{0}} is the original learning rate, d {\displaystyle d} is a decay parameter and n {\displaystyle n} is the iteration step. Step-based learning schedules changes the learning rate according to some predefined steps. The decay application formula is here defined as: η n = η 0 d ⌊ 1 + n r ⌋ {\displaystyle \eta _{n}=\eta _{0}d^{\left\lfloor {\frac {1+n}{r}}\right\rfloor }} where η n {\displaystyle \eta _{n}} is the learning rate at iteration n {\displaystyle n} , η 0 {\displaystyle \eta _{0}} is the initial learning rate, d {\displaystyle d} is how much the learning rate should change at each drop (0.5 corresponds to a halving) and r {\displaystyle r} corresponds to the drop rate, or how often the rate should be dropped (10 corresponds to a drop every 10 iterations). The floor function ( ⌊ … ⌋ {\displaystyle \lfloor \dots \rfloor } ) here drops the value of its input to 0 for all values smaller than 1. Exponential learning schedules are similar to step-based, but instead of steps, a decreasing exponential function is used. The mathematical formula for factoring in the decay is: η n = η 0 e − d n {\displaystyle \eta _{n}=\eta _{0}e^{-dn}} where d {\displaystyle d} is a decay parameter. == Adaptive learning rate == The issue with learning rate schedules is that they all depend on hyperparameters that must be manually chosen for each given learning session and may vary greatly depending on the problem at hand or the model used. To combat this, there are many different types of adaptive gradient descent algorithms such as Adagrad, Adadelta, RMSprop, and Adam which are generally built into deep learning libraries such as Keras.
Pixel binning
Pixel binning, also known as binning, is a process image sensors of digital cameras use to combine adjacent pixels throughout an image, by summing or averaging their values, during or after readout. It improves low-light performance while still allowing for highly detailed photographs in good light. Charge from adjacent pixels in CCD or charge-coupled device image sensors and some other image sensors can be combined during readout, increasing the line rate or frame rate. In the context of image processing, binning is the procedure of combining clusters of adjacent pixels, throughout an image, into single pixels. For example, in 2×2 binning, an array of 4 pixels becomes a single larger pixel, reducing the number of pixels to 1/4 and halving the image resolution in each dimension. The result can be the sum, average, median, minimum, or maximum value of the cluster. Some systems use more advanced algorithms such as considering the values of nearby pixels, edge detection, self-claimed "AI", etc. to increase the perceived visual quality of the final downsized image. This aggregation, although associated with loss of information, reduces the amount of data to be processed, facilitating analysis. The binned image has lower resolution, but the relative noise level in each pixel is generally reduced. == History == Normally, an increase in megapixel count on a constant image sensor size would lead to a sacrifice of the surface size of the individual pixels, which would result in each pixel being able to catch less light in the same time, thus leading to a darker and/or noisier image in low light (given the same exposure time). In the past, camera manufacturers had to compromise between low-light performance and the amount of detail in good light, by dropping the megapixel count like HTC did in 2013 with their four-megapixel "UltraPixel" camera. However, this results in less detailed images in daylight where enough light is available. With pixel binning, the camera has "the best of both worlds", meaning both the benefit of high detail in good light and the benefit of high brightness in low light. In low light, the surfaces of four or more pixels can act as one large pixel that catches far more light. For example, some smartphones such as the Samsung Galaxy A15 are able to capture photographs with up to fifty megapixels in daylight. However, in low light, the individual pixels would be too small to capture the light needed for a bright image with the short exposure time available for handheld shooting. Therefore, with pixel binning activated, the 50-megapixel image sensor acts as a 12.5-megapixel image sensor, a quarter of its original resolution, with an accordingly larger surface area per pixel.
Intelligent Robotics Group
The Intelligent Robotics Group (IRG) is a research organization within the Intelligent Systems Division at the NASA Ames Research Center in California's Silicon Valley. IRG conducts applied research in the area of robotics and autonomy and is one of the principal organizations at NASA responsible for robotics expertise, along with groups at the Jet Propulsion Laboratory and Johnson Space Center. The group's portfolio includes robotics in support of human exploration, perception and navigation, user interfaces, software architectures, and simulation. IRG developed the Astrobee free-flying robots on the International Space Station and was a primary contributor to the VIPER lunar rover in the areas of flight software, navigation, simulation, and mission operations. IRG has also conducted many robotic field test campaigns in support of spaceflight mission concept developments. These experiences led to the commercialization of the GigaPan system in collaboration with Carnegie Mellon University.
Strategic Computing Initiative
The United States government's Strategic Computing Initiative funded research into advanced computer hardware and artificial intelligence from 1983 to 1993. The initiative was designed to support various projects that were required to develop machine intelligence in a prescribed ten-year time frame, from chip design and manufacture, computer architecture to artificial intelligence software. The Department of Defense spent a total of $1 billion on the project. The inspiration for the program was Japan's fifth generation computer project, an enormous initiative that set aside billions for research into computing and artificial intelligence. As with Sputnik in 1957, the American government saw the Japanese project as a challenge to its technological dominance. The British government also funded a program of their own around the same time, known as Alvey, and a consortium of U.S. companies funded another similar project, the Microelectronics and Computer Technology Corporation. The goal of SCI, and other contemporary projects, was nothing less than full machine intelligence. "The machine envisioned by SC", according to Alex Roland and Philip Shiman, "would run ten billion instructions per second to see, hear, speak, and think like a human. The degree of integration required would rival that achieved by the human brain, the most complex instrument known to man." The initiative was conceived as an integrated program, similar to the Apollo moon program, where different subsystems would be created by various companies and academic projects and eventually brought together into a single integrated system. Roland and Shiman wrote that "While most research programs entail tactics or strategy, SC boasted grand strategy, a master plan for an entire campaign." The project was funded by the Defense Advanced Research Projects Agency and directed by the Information Processing Technology Office (IPTO). By 1985 it had spent $100 million, and 92 projects were underway at 60 institutions: half in industry, half in universities and government labs. Robert Kahn, who directed IPTO in those years, provided the project with its early leadership and inspiration. Clint Kelly managed the SC Initiative for three years and developed many of the specific application programs for DARPA, such as the Autonomous Land Vehicle. By the late 1980s, it was clear that the project would fall short of realizing the hoped-for levels of machine intelligence. Program insiders pointed to issues with integration, organization, and communication. When Jack Schwarz ascended to the leadership of IPTO in 1987, he cut funding to artificial intelligence research (the software component) "deeply and brutally", "eviscerating" the program (wrote Pamela McCorduck). Schwarz felt that DARPA should focus its funding only on those technologies which showed the most promise. In his words, DARPA should "surf", rather than "dog paddle", and he felt strongly AI was not "the next wave". The project was superseded in the 1990s by the Accelerated Strategic Computing Initiative and then by the Advanced Simulation and Computing Program. These later programs did not include artificial general intelligence as a goal, but instead focused on supercomputing for large scale simulation, such as atomic bomb simulations. The Strategic Computing Initiative of the 1980s is distinct from the 2015 National Strategic Computing Initiative—the two are unrelated. == Results == Although the program failed to meet its goal of high-level machine intelligence, it did meet some of its specific technical objectives, for example those of autonomous land navigation. The Autonomous Land Vehicle program and its sister Navlab project at Carnegie Mellon University, in particular, laid the scientific and technical foundation for many of the driverless vehicle programs that came after it, such as the Demo II and III programs (ALV being Demo I), Perceptor, and the DARPA Grand Challenge. The use of video cameras plus laser scanners and inertial navigation units pioneered by the SCI ALV program form the basis of almost all commercial driverless car developments today. It also helped to advance the state of the art of computer hardware to a considerable degree. On the software side, the initiative funded development of the Dynamic Analysis and Replanning Tool (DART), a program that handled logistics using artificial intelligence techniques. This was a huge success, saving the Department of Defense billions during Desert Storm. Introduced in 1991, DART had by 1995 offset the monetary equivalent of all funds DARPA had channeled into AI research for the previous 30 years combined.
Z.ai
Knowledge Atlas Technology Joint Stock Co., Ltd., branded internationally as Z.ai, is a Chinese technology company specializing in artificial intelligence (AI). The company was formerly known as Zhipu AI outside China until its rebranding in 2025. Z.ai's flagship product is the GLM (General Language Model) family of large language models, which the company has released under the free and open-source MIT License since July 2025. As of 2024, it is one of China's "AI tiger" companies by investors and considered to be the third-largest LLM market player in China's AI industry according to the International Data Corporation. In January 2025, the United States Commerce Department blacklisted the company in its Entity List due to national security concerns. == History == Founded in 2019, the startup company began from Tsinghua University and was later spun out as an independent company. Researchers published an Association for Computational Linguistics conference paper in May 2022 introducing the GLM (General Language Model) training algorithm, which uses an "autoregressive blank infilling" strategy that creates cloze tests by randomly removing segments of input text and trains the model to autoregressively regenerate the removed text. In 2023, it raised 2.5 billion yuan (approx. 350 million in USD) from Alibaba Group and Tencent, along with Meituan, Ant Group, Xiaomi, and HongShan. In March 2024, Zhipu AI announced it was developing a Sora-like technology to achieve artificial general intelligence (AGI). In May 2024, the Saudi Arabian finance firm Prosperity7 Ventures, LLC participated in a USD $400 million financing round for Zhipu AI with a valuation of approximately 3 billion USD. In July 2024, they debuted the Ying text-to-video model. Zhipu released GLM-4-Plus in August 2024. In October 2024, Zhipu released GLM-4-Voice, an end-to-end speech large language model that can adjust its tone or dialect. Zhipu disclosed in April 2025 that it had started preparing for its initial public offering (IPO) and released two models under the free and open-source MIT License. In May 2025, the company sealed a 61.28 million yuan deal from the Chinese government for city projects in Hangzhou. In July 2025, Zhipu AI released GLM-4.5 and GLM-4.5 Air, their next generation language models, and the company rebranded itself as Z.ai internationally. In August 2025, Z.ai announced that their GLM models are compatible with Huawei's Ascend processors. On August 11, 2025, Z.ai released a new vision-language model (VLM) with a total of 106B parameters, GLM-4.5V. In late September 2025, the company released GLM-4.6 using China's domestic chips such as those from Cambricon Technologies. Z.ai released GLM-4.6V and GLM-4.7 in December 2025. That same year, the company changed its official name to Knowledge Atlas Technology JSC Ltd. On 8 January 2026, Z.ai held its IPO on the Hong Kong Stock Exchange to become a listed company. It is considered to be China's first major LLM company that went through an IPO. On February 11, 2026, Z.ai released GLM-5. In late February 2026, Z.ai's shares fell by 23%, and had a shortage of compute resources, leading to user complaints and Z.ai issuing a public call for support. Z.ai also restricted new user signups. In late March, 2026, Z.ai released the GLM-5.1 model to subscription users. On April 8th, 2026, Z.ai released GLM-5.1 as open-source. The same day, Z.ai increased its API prices by 10%, but maintained a lower price than its United States competitor Anthropic's Opus 4.6 model. On release, the company's share price increased 11.5%. == Description == Z.ai provides the following products and services: General Language Model (commonly abbreviated as GLM; formerly known as ChatGLM), a series of pre-trained dialogue models initially developed by Zhipu AI and Tsinghua KEG in 2023. GLM 4.5, released in July 2025 by Z.ai, can run on eight NVIDIA H20 chips. The release of GLM-4.6 in late September 2025 marked the first integration of FP8 and Int4 quantization on Cambricon chips. It also supports native FP8 on Moore Threads GPUs. Ying, a text-to-video model that generates image and text prompts into a six-second video clip for around 30 seconds. AutoGLM, an AI agent application that uses voice commands to complete tasks within a smartphone. The app can analyze complex tasks such as ordering an item from a nearby store and repeating an order based from the user's shopping history. AMiner, created by Jie Tang (co-founder of Z.ai) in March 2006, now owned by Z.ai. Z.ai has offices in the Middle East, United Kingdom, Singapore, and Malaysia, along with innovation center projects across Southeast Asia (2025). In January 2025, the United States Commerce Department added the company to its Entity List, citing national security concerns. == Models ==
AI warfare
AI warfare refers to the use of artificial intelligence technologies to automate military operation and enhance or bypass human decision-making in armed conflicts. AI is used to rapidly analyze large volumes of military intelligence data, including making recommendations or decisions on who and what to target. Abdul-Rahman al-Rawi, a 20-year-old student, was the first acknowledged civilian killed by AI-assisted airstrike in a U.S. strike in Iraq in 2024. In 2026, the U.S. declared it would become an 'AI-first' warfighting force. Husain et al (2018) coined the term hyperwar to refer to warfare which is algorithmic or controlled by artificial intelligence, with little to no human decision-making. == 2026 Iran war == The 2026 Iran war has been described as the "first AI war", although the Untied States and Israel have previously used AI to identify targets during the Gaza war. The U.S. has used AI tools to attack Iran. These tools have been used for military intelligence, targeting, and damage assessment in the war in Iran. Using the Maven smart system, the U.S. attacked 1,000 targets in the first 24 hours of the war and 5,000 targets over the course of 10 days. While the U.S. had used Maven in 2022 to share targeting information with Ukraine and strike against Iraq, Syria, and against the Houthis in 2024, Iran's attacks are its biggest. Authorities are looking into whether artificial intelligence was involved in the airstrike on an Iranian girls' school that killed 170 civilians, the majority of whom were female students. The United States Central Command emphasized that humans were making final targeting decisions. Per a White House tally released on April 8, the U.S. military hit over 13,000 targets in Iran during the war's first 38 days, including more than 2,000 command-and-control sites, 1,500 air defense targets, and 1,450 industrial infrastructure targets. == Gaza war == As part of the Gaza war, the Israel Defense Forces (IDF) have used artificial intelligence to rapidly and automatically perform much of the process of determining what to bomb. IDF's Unit 8200 developed AI systems, dubbed the Gospel and Lavender, to find targets for the Israeli Air Force to bomb. The Gospel automatically provides targeting recommendations to human analysts, who decide whether to approve strikes. Lavender identified 37,000 Hamas-linked individuals early in the war, and was used alongside the Gospel, which chooses buildings or structures as targets. According to a report by +972 Magazine and Local Call, strikes assisted by Lavender were routinely permitted to kill 5–20 civilians for each suspected Hamas militant, who were often bombed at home with their families. The IDF denies these claims, maintaining that every strike is assessed to minimize collateral damage, and that there is no policy "to kill tens of thousands of people in their homes." Israel deployed AI technologies during the Gaza war for audio analysis, facial recognition, and airstrike targeting. One such system was used to help identify the location of Hamas commander Ibrahim Biari through phone call analysis, leading to strikes that killed him as well as more than 125 civilians. == 2022 Russian Ukraine war == Kyiv launched a project with Palantir called Brave1 Dataroom to build AI systems using the extensive combat data Ukraine has gathered since Russia’s full-scale invasion in 2022. The country has also created tools for in-depth airstrike analysis, introduced AI to process large volumes of intelligence, and incorporated these technologies into the planning of long-range strike operations. == Involved companies == Maven Smart System is developed by Palantir. It integrates Anthropic's Claude as its large language model, and uses Amazon's AWS servers as its cloud infrastructure. Since Anthropic's refusal to support autonomous weapons development and domestic surveillance efforts. In its place, other AI firms, including OpenAI, have been brought in to take over that role. == Involved state actors == In 2024, the United States Department of Defense had 800-plus active AI-related projects and requested $1.8 billion in AI funding, with Project Maven and Project Artemis (AI-resistant drones developed together with Ukraine) being the main ones. The technology has been used in Iran, Iraq, Syria and Yemen to identify targets. China is pursuing intelligentized warfare, integrating AI across all combat domains—land, sea, air, space, and cyber—with military AI spending exceeding $1.6 billion annually. == International regulation == Since 2014, states meeting within the framework of the Convention on Certain Conventional Weapons have discussed lethal autonomous weapon systems. In 2016, the treaty's states parties established an open-ended Group of Governmental Experts on Lethal Autonomous Weapons Systems to continue those discussions. The discussions have addressed international humanitarian law, accountability, possible prohibitions and regulations, and the extent of human control required over AI-enabled weapons.
RuleML
RuleML is a global initiative, led by a non-profit organization RuleML Inc., that is devoted to advancing research and industry standards design activities in the technical area of rules that are semantic and highly inter-operable. The standards design takes the form primarily of a markup language, also known as RuleML. The research activities include an annual research conference, the RuleML Symposium, also known as RuleML for short. Founded in fall 2000 by Harold Boley, Benjamin Grosof, and Said Tabet, RuleML was originally devoted purely to standards design, but then quickly branched out into the related activities of coordinating research and organizing an annual research conference starting in 2002. The M in RuleML is sometimes interpreted as standing for Markup and Modeling. The markup language was developed to express both forward (bottom-up) and backward (top-down) rules in XML for deduction, rewriting, and further inferential-transformational tasks. It is defined by the Rule Markup Initiative, an open network of individuals and groups from both industry and academia that was formed to develop a canonical Web language for rules using XML markup and transformations from and to other rule standards/systems. Markup standards and initiatives related to RuleML include: Rule Interchange Format (RIF): The design and overall purpose of W3C's Rule Interchange Format (RIF) industry standard is based primarily on the RuleML industry standards design. Like RuleML, RIF embraces a multiplicity of potentially useful rule dialects that nevertheless share common characteristics. RuleML Technical Committee from Oasis-Open: An industry standards effort devoted to legal automation utilizing RuleML. Semantic Web Rule Language (SWRL): An industry standards design, based primarily on an early version of RuleML, whose development was funded in part by the DARPA Agent Markup Language (DAML) research program. Semantic Web Services Framework, particularly its Semantic Web Services Language: An industry standards design, based primarily on a medium-mature version of RuleML, whose development was funded in part by the DARPA Agent Markup Language (DAML) research program and the WSMO research effort of the EU. Mathematical Markup Language (MathML): However, MathML's Content Markup is better suited for defining functions rather than relations or general rules Predictive Model Markup Language (PMML): With this XML-based language one can define and share various models for data-mining results, including association rules Attribute Grammars in XML (AG-markup): For AG's semantic rules, there are various possible XML markups that are similar to Horn-rule markup Extensible Stylesheet Language Transformations (XSLT): This is a restricted term-rewriting system of rules, written in XML, for transforming XML documents into other text documents