The Transparency in Frontier Artificial Intelligence Act, also referred to as SB-53, is a 2025 California law which mandates increased transparency for companies building artificial intelligence. SB-53 is primarily focused on assessing and reducing potential catastrophic risks from AI, and is the first bill addressing such risks to be passed into law in America. The bill requires companies to create publicly accessible documents assessing potential "catastrophic risk[s]" from their AI models, as well as publishing documentation on how the model incorporates national and international safety standards. SB-53 also sets up whistleblower protections and procedures for alerting the government to a "critical safety incident". == History == SB-53 was preceded in 2024 by the unsuccessful Safe and Secure Innovation for Frontier Artificial Intelligence Models Act ("SB-1047"), a proposed bill authored by Senator Scott Wiener which was vetoed by Governor Gavin Newsom. Afterwords, Newsom created a "Joint California AI Policy Working Group" to provide recommendations for AI regulation, which guided the drafting of SB-53. Senator Scott Wiener introduced the bill on January 7, 2025, and after a series of amendments, SB-53 passed the Senate 29-8 on September 13. Governor Gavin Newsom approved the bill on September 25, passing it into law. == Provisions == SB-53 applies primarily to companies making at least $500 million in yearly gross revenue. It defines a “frontier model” as any AI trained with over 1026 FLOPS (including fine-tuning), including unreleased internal models. Both the financial and computational thresholds must be met before most of the law is applied, although the threshold can be lowered or otherwise updated by the California Department of Technology in an annual review starting in 2027. Most of the bill's provisions are focused on "catastrophic risks" from AI, which are defined as incidents in which a model contributes to more than 50 deaths or serious injuries, or causes more than one billion dollars ($1,000,000,000) in economic damage from AI-assisted acts (such as cyberattacks or the creation of biological weapons). The bill requires companies to provide publicly accessible safety frameworks for frontier AI models, describing how the company tests for catastrophic risk from its AI, and how it implements protections against such risks. This includes addressing the possibility that the AI may attempt to circumvent internal guardrails or oversight mechanisms. (Certain safety incidents, such as dangerously deceptive model behavior, physical injury, or death, must be reported to California Office of Emergency Services (OES) within 15 days, unless the incident poses imminent physical risk, in which case it must be reported immediately.) The company must follow its published framework, and if any changes are made, the framework should be updated within 30 days, and justification for said changes must also be made public. Additionally, all frontier companies are required to publish basic information about newly released frontier models (such as terms of service, supported languages, and intended use), although only large companies (making over $500 million annually) need to publish full safety frameworks. SB-53 also establishes various whistleblower protections for covered employees. Large companies must have anonymous whistleblowing channels in place which protect employees from retaliation from reporting risks to state or federal authorities if they have reasonable cause to believe that their employer is substantially risking public health and safety.
Stochastic parrot
In machine learning, the term stochastic parrot is a metaphor that frames large language models as systems that statistically mimic text without real understanding. The word "stochastic" – from the ancient Greek "στοχαστικός" (stokhastikos, 'based on guesswork') – is a term from probability theory meaning "randomly determined". The word "parrot" refers to parrots' ability to mimic human speech. The term was introduced in a 2021 paper on AI ethics titled "On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜" and authored by Timnit Gebru, Emily M. Bender, Angelina McMillan-Major, and Margaret Mitchell. The paper outlined possible risks associated with large language models (LLMs). In December 2020, it was the subject of a workplace dispute between Gebru (then co-leader of Google's Ethical Artificial Intelligence Team) and Google, which had requested the retraction of the paper. The incident culminated in Gebru's controversial departure from the company. The paper was later presented at the 2021 ACM Conference, and the term "stochastic parrot" has seen widespread use in academic research concerning generative AI and LLMs. The term has been interpreted negatively as an insult towards AI. == Background == Timnit Gebru is an AI ethics researcher, Emily M. Bender is a linguist specializing in computational linguistics, and Margaret Mitchell is a computer scientist specializing in algorithmic bias. Gebru had joined Google in 2018, where she co-led a team on the ethics of artificial intelligence with Mitchell. In late 2020, the paper "On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜" was co-written by Gebru and five other researchers, four of whom were Google employees. The paper argues that large language models (LLMs) present significant risks such as environmental and financial costs, inscrutability leading to unknown dangerous biases, and potential for deception as LLMs do not understand the concepts underlying what they learn. The paper states that LLMs are "stitching together sequences of linguistic forms ... observed in its vast training data, according to probabilistic information about how they combine, but without any reference to meaning." Therefore, they are labeled "stochastic parrots". === Dismissal of Gebru by Google === After the paper was submitted for consideration to the 2021 ACM Conference, Google requested that Gebru either retract the paper from the conference or remove the names of Google employees from it. Gebru refused to do so without further discussion, and emailed Google Research vice president Megan Kacholia that if the company could not explain the request for retraction and address other concerns regarding similar projects, she would plan to resign after a transition period, stating that they could "work on a last date". The following day, on December 2, 2020, Gebru received an email saying that Google was "accepting her resignation". Her abrupt firing sparked protests by Google employees and negative publicity for the company. == Usage == The phrase has been used by AI skeptics to signify that LLMs lack understanding of the meaning of their outputs. Sam Altman, CEO of OpenAI, used the term shortly after the release of ChatGPT in December 2022, tweeting "i am a stochastic parrot, and so r u". The term was nominated as the 2023 AI-related Word of the Year by the American Dialect Society. == Debate == Some LLMs, such as ChatGPT, have become capable of interacting with users in convincingly human-like conversations. The development of these new systems has deepened the discussion of the extent to which LLMs understand or are simply "parroting". According to machine learning researchers Lindholm, Wahlström, Lindsten, and Schön, the term "stochastic parrot" highlights two vital limitations of LLMs: LLMs are limited by the data they are trained on and are simply stochastically repeating contents of datasets. Because they are just making up outputs based on training data, LLMs do not understand if they are saying something incorrect or inappropriate. Lindholm et al. noted that, with poor quality datasets and other limitations, a learning machine might produce results that are "dangerously wrong". === Subjective experience === In the mind of a human being, words and language correspond to things one has experienced. For LLMs, according to proponents of the theory, words correspond only to other words and patterns of usage fed into their training data. Proponents of the idea of stochastic parrots thus conclude that statements about LLMs are due to "the human tendency to attribute meaning to text", and claim this occurs despite the LLMs not actually understanding language. === Fine-tuning === Kelsey Piper argued that the claim that LLMs are stochastic parrots or mere "next-token predictors" focuses on pre-training, ignoring that modern LLMs are also fine-tuned to follow instructions and to prefer accurate answers. === Hallucinations and mistakes === The tendency of LLMs to pass off false information as fact is held as support. Called hallucinations or confabulations, LLMs will occasionally synthesize information that matches some pattern. LLMs may fail to distinguish fact and fiction, which leads to the claim that they can't connect words to a comprehension of the world, as humans do. Furthermore, LLMs may fail to decipher complex or ambiguous grammar cases that rely on understanding the meaning of language. For example: The wet newspaper that fell down off the table is my favorite newspaper. But now that my favorite newspaper fired the editor I might not like reading it anymore. Can I replace 'my favorite newspaper' by 'the wet newspaper that fell down off the table' in the second sentence? GPT-4, an LLM released in March 2023, responded yes, not understanding that the meaning of "newspaper" is different in these two contexts; it is first an object and second an institution. === Benchmarks and experiments === One argument against the hypothesis that LLMs are stochastic parrot is their results on benchmarks for reasoning, common sense and language understanding. In 2023, some LLMs have shown good results on many language understanding tests, such as the Super General Language Understanding Evaluation (SuperGLUE). GPT-4 scored in the >90th-percentile on the Uniform Bar Examination and achieved 93% accuracy on the MATH benchmark of high-school Olympiad problems, results that exceed rote pattern-matching expectations. Such tests, and the smoothness of many LLM responses, help as many as 51% of AI professionals believe they can truly understand language with enough data, according to a 2022 survey. === Expert rebuttals === Some AI researchers dispute the notion that LLMs merely "parrot" their training data. Geoffrey Hinton, a pioneering figure in neural networks, counters that the metaphor misunderstands the prerequisite for accurate language prediction. He argues that "to predict the next word accurately, you have to understand the sentence", a view he presented on 60 Minutes in 2023. From this perspective, understanding is not an alternative to statistical prediction, but rather an emergent property required to perform it effectively at scale. Hinton also uses logical puzzles to demonstrate that LLMs actually understand language. A 2024 Scientific American investigation described a closed Berkeley workshop where state-of-the-art models solved novel tier-4 mathematics problems and produced coherent proofs, indicating reasoning abilities beyond memorization. The GPT-4 Technical Report showed human-level results on professional and academic exams (e.g., the Uniform Bar Exam and USMLE), challenging the "parrot" characterization. Anthropic conducted mechanistic interpretability research on Claude, using attribution graphs to identify circuits. The research showed how the LLM processes information via chains of fuzzy logical inference, and indicated an ability to plan ahead. They found that Claude 3.5 Haiku "employs remarkably general abstractions", forms "internally generated plans for its future outputs" and "works backwards from its longer-term goals". They noted that "The mechanisms of the model can apparently only be faithfully described using an overwhelmingly large causal graph." They also found that the model includes "mechanisms that could underlie a simple form of metacognition", in that it "thinks about" the level of its own knowledge before reaching its answer. === Interpretability === Another line of evidence against the 'stochastic parrot' claim comes from mechanistic interpretability, a research field dedicated to reverse-engineering LLMs to understand their internal workings. Rather than only observing the model's input-output behavior, these techniques probe the model's internal activations, which can be used to determine if they contain structured representations of the world. The goal is to investigate whether LLMs are merely manipulating surface statistics or if t
Pedagogical agent
A pedagogical agent is a concept borrowed from computer science and artificial intelligence and applied to education, usually as part of an intelligent tutoring system (ITS). It is a simulated human-like interface between the learner and the content, in an educational environment. A pedagogical agent is designed to model the type of interactions between a student and another person. Mabanza and de Wet define it as "a character enacted by a computer that interacts with the user in a socially engaging manner". A pedagogical agent can be assigned different roles in the learning environment, such as tutor or co-learner, depending on the desired purpose of the agent. "A tutor agent plays the role of a teacher, while a co-learner agent plays the role of a learning companion". == History == The history of Pedagogical Agents is closely aligned with the history of computer animation. As computer animation progressed, it was adopted by educators to enhance computerized learning by including a lifelike interface between the program and the learner. The first versions of a pedagogical agent were more cartoon than person, like Microsoft's Clippy which helped users of Microsoft Office load and use the program's features in 1997. However, with developments in computer animation, pedagogical agents can now look lifelike. By 2006 there was a call to develop modular, reusable agents to decrease the time and expertise required to create a pedagogical agent. There was also a call in 2009 to enact agent standards. The standardization and re-usability of pedagogical agents is less of an issue since the decrease in cost and widespread availability of animation tools. Individualized pedagogical agents can be found across disciplines including medicine, math, law, language learning, automotive, and armed forces. They are used in applications directed to every age, from preschool to adult. == Learning theories related to pedagogical agent design == === Distributed cognition theory === Distributed cognition theory is the method in which cognition progresses in the context of collaboration with others. Pedagogical agents can be designed to assist the cognitive transfer to the learner, operating as artifacts or partners with collaborative role in learning. To support the performance of an action by the user, the pedagogical agent can act as a cognitive tool as long as the agent is equipped with the knowledge that the user lacks. The interactions between the user and the pedagogical agent can facilitate a social relationship. The pedagogical agent may fulfill the role of a working partner. === Socio-cultural learning theory === Socio-cultural learning theory is how the user develops when they are involved in learning activities in which there is interaction with other agents. A pedagogical agent can: intervene when the user requests, provide support for tasks that the user cannot address, and potentially extend the learners cognitive reach. Interaction with the pedagogical agent may elicit a variety of emotions from the learner. The learner may become excited, confused, frustrated, and/or discouraged. These emotions affect the learners' motivation. === Extraneous Cognitive Load === Extraneous cognitive load is the extra effort being exerted by an individual's working memory due to the way information is being presented. A pedagogical agent can increase the user's cognitive load by distracting them and becoming the focus of their attention, causing split attention between the instructional material and the agent. Agents can reduce the perceived cognitive load by providing narration and personalization that can also promote a user's interest and motivation. While research on the reduction of cognitive load from pedagogical agents is minimal, more studies have shown that agents do not increase it. == Effectiveness == It has been suggested by researchers that pedagogical agents may take on different roles in the learning environment. Examples of these roles are: supplanting, scaffolding, coaching, testing, or demonstrating or modelling a procedure. A pedagogical agent as a tutor has not been demonstrated to add any benefit to an educational strategy in equivalent lessons with and without a pedagogical agent. According to Richard Mayer, there is some support in research for pedagogical agent increasing learning, but only as a presenter of social cues. A co-learner pedagogical agent is believed to increase the student's self-efficacy. By pointing out important features of instructional content, a pedagogical agent can fulfill the signaling function, which research on multimedia learning has shown to enhance learning. Research has demonstrated that human-human interaction may not be completely replaced by pedagogical agents, but learners may prefer the agents to non-agent multimedia systems. This finding is supported by social agency theory. Much like the varying effectiveness of the pedagogical agent roles in the learning environment, agents that take into account the user's affect have had mixed results. Research has shown pedagogical agents that make use of the users’ affect have been found to increase user knowledge retention, motivation, and perceived self-efficacy. However, with such a broad range of modalities in affective expressions, it is often difficult to utilize them. Additionally, having agents detect a user's affective state with precision remains challenging, as displays of affect are different across individuals. == Design == === Attractiveness === The appearance of a pedagogical agent can be manipulated to meet the learning requirements. The attractiveness of a pedagogical agent can enhance student's learning when the users were the opposite gender of the pedagogical agent. Male students prefer a sexy appearance of a female pedagogical agents and dislike the sexy appearance of male agents. Female students were not attracted by the sexy appearance of either male or female pedagogical agents. === Affective Response === Pedagogical agents have reached a point where they can convey and elicit emotion, but also reason about and respond to it. These agents are often designed to elicit and respond to affective actions from users through various modalities such as speech, facial expressions, and body gestures. They respond to the affective state of the given user, and make use of these modalities using a wide array of sensors incorporated into the design of the agent. Specifically in education and training applications, pedagogical agents are often designed to increasingly recognize when users or learners exhibit frustration, boredom, confusion, and states of flow. The added recognition in these agents is a step toward making them more emotionally intelligent, comforting and motivating the users as they interact. === Digital Representation === The design of a pedagogical agent often begins with its digital representation, whether it will be 2D or 3D and static or animated. Several studies have developed pedagogical agents that were both static and animated, then evaluated the relative benefits. Similar to other design considerations, the improved learning from static or animated agents remains questionable. One study showed that the appearance of an agent portrayed using a static image can impact a user's recall, based on the visual appearance. Other research found results that suggest static agent images improve learning outcomes. However, several other studies found user's learned more when the pedagogical agent was animated rather than static. Recently a meta-analysis of such research found a negligible improvement in learning via pedagogical agents, suggesting more work needs to be done in the area to support any claims.
Artificial intelligence and elections
As artificial intelligence (AI) has become more mainstream, there is growing concern about how this will influence elections. Potential targets of AI include election processes, election offices, election officials and election vendors. There are also global efforts to improve elections using AI. == Tactics == Generative AI capabilities allow creation of misleading content. Examples of this include text-to-video, deepfake videos, text-to-image, AI-altered images, text-to-speech, voice cloning, and text-to-text. In the context of an election, a deepfake video of a candidate may propagate information that the candidate does not endorse. Chatbots could spread misinformation related to election locations, times or voting methods. In contrast to malicious actors in the past, these techniques require little technical skill and can spread rapidly. LLM-generated messages have the capacity to persuade humans on political issues. Researchers have begun to investigate how people rate messages that LLMs generate for how persuasive they are. When it came to policy issues, the LLM-generated messages received a 2.91 compared to a 2.80 when it came to smartness between the AI and humans. The LLM-generated messages were often more technical and analytical than human-generated messages. Generative AI has been used to micro-target people during tight political elections. The generation of targeted large language models has triggered concern that they will be used to leverage readily scale microtargeting. Rephrasing inputs have been used to generate fraudulent emails and phishing websites. Rephrasing inputs in a microtargeting does not violate the terms of OpenAI usage. There are no safeguards to prevent the use of rephrasing and creation of fraudulent emails. Political campaign managers have access to this allowing for them to create targeted content. == Usage by country == === Argentina === ==== 2023 elections ==== During the 2023 Argentine primary elections, Javier Milei's team distributed AI generated images including a fabricated image of his rival Sergio Massa and drew 3 million views. The team also created an unofficial Instagram account entitled "AI for the Homeland." Sergio Massa's team also distributed AI generated images and videos. === Bangladesh === ==== 2024 elections ==== In the run up to the 2024 Bangladeshi general election, deepfake videos of female opposition politicians appeared. Rumin Farhana was pictured in a bikini while Nipun Ray was shown in a swimming pool. === Canada === ==== 2025 elections ==== In the run up to the 2025 Canadian federal election, the use of AI tools is likely to figure prominently. India, Pakistan and Iran are all expected to make efforts to subvert the national vote using disinformation campaigns to deceive voters and sway diaspora communities. In a report by the Canadian Centre for Cyber Security called "Cyber Threats to Canada's Democratic Process: 2025 Update", it states that malicious actors including China and Russia: "are most likely to use generative AI as a means of creating and spreading disinformation, designed to sow division among Canadians and push narratives conducive to the interests of foreign states". === France === ==== 2024 elections ==== In the 2024 French legislative election, deepfake videos appeared claiming: i) That they showed the family of Marine le Pen. In the videos, young women, supposedly Le Pen's nieces, are seen skiing, dancing and at the beach "while making fun of France’s racial minorities": However, the family members don't exist. On social media there were over 2 million views. ii) In a video seen on social media, a deepfake video of a France24 broadcast appeared to report that the Ukrainian leadership had "tried to lure French president Emmanuel Macron to Ukraine to assassinate him and then blame his death on Russia". === Ghana === ==== 2024 elections ==== During the months before the December 2024 Ghanaian general election, a network of at least 171 fake accounts has been used to spam social media. Posts have been used by a group identified as "@TheTPatriots" to promote the New Patriotic Party, although it is not known whether the two are connected. All the networks' posts were "highly likely" to have been generated by ChatGPT and appear to be the "first secretly partisan network using AI to influence elections in Ghana". The opposition National Democratic Congress was also criticized with its leader John Mahama being called a drunkard. === India === ==== 2024 elections ==== In the 2024 Indian general election, politicians used deepfakes in their campaign materials. These deepfakes included politicians who had died prior to the election. Mathuvel Karunanidhi's party posted with his likeness even though he had died 2018. A video The All-India Anna Dravidian Progressive Federation party posted showed an audio clip of Jayaram Jayalalithaa even though she had died in 2016. The Deepfakes Analysis Unit (DAU) is an open source platform created in March 2024 for the public to share misleading content and assess if it had been AI-generated. AI was also used to translate political speeches in real time. This translating ability was widely used to reach more voters. === Indonesia === ==== 2024 elections ==== In the 2024 Indonesian presidential election, Prabowo Subianto made extensive use of AI-generated art in his campaign, which ranged from images of himself as an adorable child to various child portrayals in his advertisements. The Indonesian Children's Protection Commission condemned these ads, labeling them as a form of misuse. Other candidates, Anies Baswedan and Ganjar Pranowo, also incorporated AI art into their campaigns. Throughout the election period, all presidential candidates faced attacks from deepfakes, both in video and audio formats. === Ireland === ==== 2024 elections ==== In the last weeks of the 2024 Irish general election a spoof election poster appeared in Dublin featuring "an AI-generated candidate with three arms". The candidate is called Aidan Irwin, but no-one stood in the election with that name. A slogan on the poster says "put matters into artificial intelligence’s hands". The convincing election poster shows a man that "has six fingers on one hand, three arms, and a distorted thumb". === New Zealand === ==== 2023 elections ==== In May 2023, ahead of the 2023 New Zealand general election in October 2023, the New Zealand National Party published a "series of AI-generated political advertisements" on its Instagram account. After confirming that the images were faked, a party spokesperson said that it was "an innovative way to drive our social media". === Pakistan === ==== 2024 elections ==== AI has been used by the imprisoned ex-Prime Minister Imran Khan and his media team in the 2024 Pakistani general election: i) An AI generated audio of his voice was added to a video clip and was broadcast at a virtual rally. ii) An op-ed in The Economist written by Khan was later claimed by himself to have been written by AI which was later denied by his team. The article was liked and shared on social media by thousands of users. === South Africa === ==== 2024 elections ==== In the 2024 South African general election, there were several uses of AI content: i) A deepfaked video of Joe Biden emerged on social media showing him saying that "The U.S. would place sanctions on SA and declare it an enemy state if the African National Congress (ANC) won". ii) In a deepfake video, Donald Trump was shown endorsing the uMkhonto weSizwe party. It was posted to social media and was viewed more than 158,000 times. iii) Less than 3 months before the elections, a deepfake video showed U.S. rapper Eminem endorsing the Economic Freedom Fighters party while criticizing the ANC. The deepfake was viewed on social media more than 173,000 times. === South Korea === ==== 2022 elections ==== In the 2022 South Korean presidential election, a committee for one presidential candidate Yoon Suk Yeol released an AI avatar 'Al Yoon Seok-yeol' that would campaign in places the candidate could not go. The other presidential candidate Lee Jae-myung introduced a chatbot that provided information about the candidate's pledges. ==== 2024 elections ==== Deepfakes were used to spread misinformation before the 2024 South Korean legislative election with one source reporting 129 deepfake violations of election laws within a two week period. Seoul hosted the 2024 Summit for Democracy, a virtual gathering of world leaders initiated by US President Joe Biden in 2021. The focus of the summit was on digital threats to democracy including artificial intelligence and deepfakes. === Taiwan === ==== 2024 elections ==== AI-generated content was used during the 2024 Taiwanese presidential election. Among the media were: i) A deepfake video of General Secretary of the Chinese Communist Party Xi Jinping which showed him supporting the presidential elections. Created on social media, the video was "widely circulated
Learning rate
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.
PNGOUT
PNGOUT is a freeware command line optimizer for PNG images written by Ken Silverman. The transformation is lossless, meaning that the resulting image is visually identical to the source image. According to its author, this program can often get higher compression than other optimizers by 5–10%. It is possible to compress some inflated PNGs to a size below 1% of the original file. PNGOUT was also available as a plug-in for the freeware image viewer IrfanView and can be enabled as an option when saving files. It allows editing of various PNGOUT settings via a dialog box. PNGOUT integration was removed in IrfanView version 4.58 in favour of OptiPNG. In 2006, a commercial version of PNGOUT with a graphical user interface, known as PNGOUTWin, was released by Ardfry Imaging, a small company Silverman co-founded in 2005. There is also a freeware GUI frontend to PNGOUT available, known as PNGGauntlet. == Main operation == The main function of PNGOUT is to reduce the size of image data contained in the IDAT chunk. This chunk is compressed using the deflate algorithm. Deflate algorithms can vary in speed and compression ratio, with higher compression ratios generally implying lower speed. Ken Silverman wrote a deflate compressor for PNGOUT that is slower than the ones used in most graphics software, but produces smaller files. PNGOUT also performs automatic bit depth, color, and palette reduction where appropriate.
AlphaChip (controversy)
The AlphaChip controversy refers to a series of public, scholarly, and legal disputes surrounding a 2021 Nature paper by Google-affiliated researchers. The paper describes an approach to macro placement, a stage of chip floorplanning, based on reinforcement learning (RL), a machine learning method in which a system iteratively improves its decisions by optimizing performance-based reward signals. The primary technical question is whether the new techniques are better than existing (non-AI) techniques. Both internal Google studies and external attempts to replicate the algorithm have failed to show the claimed benefits. No head-to-head comparison is available because the data used in the paper is proprietary, and Google has not released any results from running its algorithm on public benchmarks. This has resulted in considerable skepticism over the paper's claims. In addition, the inability of others (both inside and outside of Google) to replicate the claimed results have sparked concerns about the paper’s methodology, reproducibility, and scientific integrity. The lead researchers of the Nature paper were affiliated with Google Brain, which became part of Google DeepMind, and later spun off into the company Ricursive. == Motivation for research: Macro placement in chip layout == Chip design for modern integrated circuits is a complex, expert-driven process that relies on electronic design automation. It determines the performance of the final chip, and takes weeks or months to complete. Advances that produce better designs, or complete the process faster, are commercially and academically significant. Macro placement is a step during chip design that determines the locations of large circuit components (macros) within a chip. It is followed by detailed placement, which places the far more numerous but much smaller standard cells. Alternatively, mixed-size placement simultaneously places both large macros and millions of small cells, requiring algorithms to handle objects that differ by several orders of magnitude in area and mobility. The number of macros per circuit typically ranges from several to thousands. Wiring must be performed after placement, and the details of this wiring strongly influence the power, performance, and area (PPA) of the completed chip. The full wiring calculation is very resource intensive, so placement tools typically use a proxy cost, a simplified objective function used to guide the placement algorithm during training and evaluation. The faithfulness of the chosen proxy cost to the final objective cost is a critical aspect of placer performance. === State of the art as of 2021 === Chips have been designed since the 1960s, so there were many existing methods as of 2021. Available options included manual design, academic tools, and commercial offerings. Academic methods include combinatorial optimization techniques such as simulated annealing, analytical placement, hierarchical heuristics, and as of 2019 reinforcement learning and broader machine learning techniques.. Existing (non-AI) academic tools for solving the same problem include APlace, NTUplace3, ePlace, RePlace, and DREAMPlace. Commercial EDA vendors also offered automated software tools for floorplanning and mixed-size placement. For instance, as of 2019 Cadence’s Innovus implementation software offered a Concurrent Macro Placer (CMP) feature to automatically place large blocks and standard cells. == The 2021 Nature paper and its claims == In 2021, Nature published a paper under the title “A graph‑placement methodology for fast chip design” co‑authored by 21 Google-affiliated researchers. The paper reported that an RL agent could generate macro placements for integrated circuits "in under six hours" and achieve improvements over human-designed layouts in power, timing performance, and area (PPA), standard chip-quality metrics referring respectively to energy consumption, chip operating speed, and silicon footprint (evaluated after wire routing). It introduced a sequential macro placement algorithm in which macros are placed one at a time instead of optimizing their locations concurrently. At each step, the algorithm selects a location for a single macro on a discretized chip canvas, conditioning its decision on the placements of previously placed macros. This sequential formulation converts macro placement into a long-horizon decision process in which early placement choices constrain later ones. After macro placement, force-directed placement is applied to place standard cells connected to the macros. Deep reinforcement learning is used to train a policy network to place macros by maximizing a reward that reflects final placement quality (for example, wirelength and congestion). Policy learning occurs during self‑play for one or multiple circuit designs. Further placement optimizations refine the overall layout by balancing wirelength, density, and overlap constraints, while treating the macro locations produced by the RL policy as fixed obstacles. The approach relies on pre-training, in which the RL model is first trained on a corpus of prior designs (twenty in the Nature paper) to learn general placement patterns before being fine-tuned on a specific chip. Circuit examples used in the study were parts of proprietary Google TPU designs, called blocks (or floorplan partitions). The paper reported results on five blocks and described the approach as generalizable across chip designs. == Controversy == Soon after the paper's publication, controversy arose over whether the claims were true, whether they were sufficiently proven, and whether academic standards were followed. These controversies arose both within Google and among external academic experts. === Internal dispute at Google and legal proceedings === In 2022, Satrajit Chatterjee, a Google engineer involved in reviewing the AlphaChip work, raised concerns internally and drafted an alternative analysis, (Stronger Baselines) arguing that established methods outperformed the RL approach under fair comparison. In March 2022, Google declined to publish this analysis and terminated Chatterjee's employment. Chatterjee filed a wrongful dismissal lawsuit, alleging that representations related to the AlphaChip research involved fraud and scientific misconduct. According to court documents, Chatterjee's study was conducted "in the context of a large potential Google Cloud deal". He noted that it "would have been unethical to imply that we had revolutionary technology when our tests showed otherwise" and claimed Google was deliberately withholding material information. Furthermore, the committee that reviewed his paper and disapproved its publication was allegedly chaired by subordinates of Jeff Dean, a senior co-author of the Nature paper. Google’s subsequent motion to dismiss was denied, holding that Chatterjee had plausibly alleged retaliation for refusing to engage in conduct he believed would violate state or federal law. === External controversy === The external questions can be summarized in four main points: (a) Are the claims supported by the evidence provided? (b) Did the paper provide enough information to allow the results to be independently reproduced and verified? If so, are the results an improvement over existing academic and commercial tools? (c) Were the comparisons in the paper done fairly and with full disclosure? (d) Were academic standards followed? Each of these is discussed below. ==== Are the claims supported by the evidence provided? ==== The Nature paper described the reduction in design-process time as going from "days or weeks" to "hours", but did not provide per-design time breakdowns or specify the number of engineers, their level of expertise, or the baseline tools and workflow against which this comparison was made. It was also unclear whether the "days or weeks" baseline included time spent on other tasks such as functional design changes. The paper also evaluated the method on fewer benchmarks (five) than is common in the field, and showed mixed results across different evaluation goals While the approach was described as improving circuit area, this claim seems unsupported, as the RL optimization did not alter the overall circuit area, as it adjusted only the locations of fixed-shape non-overlapping circuit components within a fixed rectangular layout boundary. ==== Comparison with existing methods, and replicating the algorithm ==== Because macro placement is largely geometric and its fundamental algorithms are not tied to a specific process node, competing approaches can be evaluated on public benchmarks (tests) across technologies, rather than primarily on proprietary internal designs. This is standard procedure when comparing academic placers, see . In contrast, Google has only reported results only on internal proprietary designs, and as of 2026 has not offered comparisons with prior methods on common benchmarks. Researchers at the University of Califor