AI Generator Question Paper

AI Generator Question Paper — independent reviews, comparisons, pricing and step-by-step guides on Aizhi.

Exposure Notification

The (Google/Apple) Exposure Notification System (GAEN) is a framework and protocol specification developed by Apple Inc. and Google to facilitate digital contact tracing during the COVID-19 pandemic. When used by health authorities, it augments more traditional contact tracing techniques by automatically logging close approaches among notification system users using Android or iOS smartphones. Exposure Notification is a decentralized reporting protocol built on a combination of Bluetooth Low Energy technology and privacy-preserving cryptography. It is an opt-in feature within COVID-19 apps developed and published by authorized health authorities. Unveiled on April 10, 2020, it was made available on iOS on May 20, 2020, as part of the iOS 13.5 update and on December 14, 2020, as part of the iOS 12.5 update for older iPhones. On Android, it was added to devices via a Google Play Services update, supporting all versions since Android Marshmallow. The Apple/Google protocol is similar to the Decentralized Privacy-Preserving Proximity Tracing (DP-3T) protocol created by the European DP-3T consortium and the Temporary Contact Number (TCN) protocol by Covid Watch, but is implemented at the operating system level, which allows for more efficient operation as a background process. Since May 2020, a variant of the DP-3T protocol is supported by the Exposure Notification Interface. Other protocols are constrained in operation because they are not privileged over normal apps. This leads to issues, particularly on iOS devices where digital contact tracing apps running in the background experience significantly degraded performance. The joint approach is also designed to maintain interoperability between Android and iOS devices, which constitute nearly all of the market. The ACLU stated the approach "appears to mitigate the worst privacy and centralization risks, but there is still room for improvement". In late April, Google and Apple shifted the emphasis of the naming of the system, describing it as an "exposure notification service", rather than "contact tracing" system. == Technical specification == Digital contact tracing protocols typically have two major responsibilities: encounter logging and infection reporting. Exposure Notification only involves encounter logging which is a decentralized architecture. The majority of infection reporting is centralized in individual app implementations. To handle encounter logging, the system uses Bluetooth Low Energy to send tracking messages to nearby devices running the protocol to discover encounters with other people. The tracking messages contain unique identifiers that are encrypted with a secret daily key held by the sending device. These identifiers change every 15–20 minutes as well as Bluetooth MAC address in order to prevent tracking of clients by malicious third parties through observing static identifiers over time. The sender's daily encryption keys are generated using a random number generator. Devices record received messages, retaining them locally for 14 days. If a user tests positive for infection, the last 14 days of their daily encryption keys can be uploaded to a central server, where it is then broadcast to all devices on the network. The method through which daily encryption keys are transmitted to the central server and broadcast is defined by individual app developers. The Google-developed reference implementation calls for a health official to request a one-time verification code (VC) from a verification server, which the user enters into the encounter logging app. This causes the app to obtain a cryptographically signed certificate, which is used to authorize the submission of keys to the central reporting server. The received keys are then provided to the protocol, where each client individually searches for matches in their local encounter history. If a match meeting certain risk parameters is found, the app notifies the user of potential exposure to the infection. Google and Apple intend to use the received signal strength (RSSI) of the beacon messages as a source to infer proximity. RSSI and other signal metadata will also be encrypted to resist deanonymization attacks. === Version 1.0 === To generate encounter identifiers, first a persistent 32-byte private Tracing Key ( t k {\displaystyle tk} ) is generated by a client. From this a 16 byte Daily Tracing Key is derived using the algorithm d t k i = H K D F ( t k , N U L L , 'CT-DTK' | | D i , 16 ) {\displaystyle dtk_{i}=HKDF(tk,NULL,{\text{'CT-DTK'}}||D_{i},16)} , where H K D F ( Key, Salt, Data, OutputLength ) {\displaystyle HKDF({\text{Key, Salt, Data, OutputLength}})} is a HKDF function using SHA-256, and D i {\displaystyle D_{i}} is the day number for the 24-hour window the broadcast is in starting from Unix Epoch Time. These generated keys are later sent to the central reporting server should a user become infected. From the daily tracing key a 16-byte temporary Rolling Proximity Identifier is generated every 10 minutes with the algorithm R P I i , j = Truncate ( H M A C ( d t k i , 'CT-RPI' | | T I N j ) , 16 ) {\displaystyle RPI_{i,j}={\text{Truncate}}(HMAC(dtk_{i},{\text{'CT-RPI'}}||TIN_{j}),16)} , where H M A C ( Key, Data ) {\displaystyle HMAC({\text{Key, Data}})} is a HMAC function using SHA-256, and T I N j {\displaystyle TIN_{j}} is the time interval number, representing a unique index for every 10 minute period in a 24-hour day. The Truncate function returns the first 16 bytes of the HMAC value. When two clients come within proximity of each other they exchange and locally store the current R P I i , j {\displaystyle RPI_{i,j}} as the encounter identifier. Once a registered health authority has confirmed the infection of a user, the user's Daily Tracing Key for the past 14 days is uploaded to the central reporting server. Clients then download this report and individually recalculate every Rolling Proximity Identifier used in the report period, matching it against the user's local encounter log. If a matching entry is found, then contact has been established and the app presents a notification to the user warning them of potential infection. === Version 1.1 === Unlike version 1.0 of the protocol, version 1.1 does not use a persistent tracing key, rather every day a new random 16-byte Temporary Exposure Key ( t e k i {\displaystyle tek_{i}} ) is generated. This is analogous to the daily tracing key from version 1.0. Here i {\displaystyle i} denotes the time is discretized in 10 minute intervals starting from Unix Epoch Time. From this two 128-bit keys are calculated, the Rolling Proximity Identifier Key ( R P I K i {\displaystyle RPIK_{i}} ) and the Associated Encrypted Metadata Key ( A E M K i {\displaystyle AEMK_{i}} ). R P I K i {\displaystyle RPIK_{i}} is calculated with the algorithm R P I K i = H K D F ( t e k i , N U L L , 'EN-RPIK' , 16 ) {\displaystyle RPIK_{i}=HKDF(tek_{i},NULL,{\text{'EN-RPIK'}},16)} , and A E M K i {\displaystyle AEMK_{i}} using the algorithm A E M K i = H K D F ( t e k i , N U L L , 'EN-AEMK' , 16 ) {\displaystyle AEMK_{i}=HKDF(tek_{i},NULL,{\text{'EN-AEMK'}},16)} . From these values a temporary Rolling Proximity Identifier ( R P I i , j {\displaystyle RPI_{i,j}} ) is generated every time the BLE MAC address changes, roughly every 15–20 minutes. The following algorithm is used: R P I i , j = A E S 128 ( R P I K i , 'EN-RPI' | | 0 x 000000000000 | | E N I N j ) {\displaystyle RPI_{i,j}=AES128(RPIK_{i},{\text{'EN-RPI'}}||{\mathtt {0x000000000000}}||ENIN_{j})} , where A E S 128 ( Key, Data ) {\displaystyle AES128({\text{Key, Data}})} is an AES cryptography function with a 128-bit key, the data is one 16-byte block, j {\displaystyle j} denotes the Unix Epoch Time at the moment the roll occurs, and E N I N j {\displaystyle ENIN_{j}} is the corresponding 10-minute interval number. Next, additional Associated Encrypted Metadata is encrypted. What the metadata represents is not specified, likely to allow the later expansion of the protocol. The following algorithm is used: Associated Encrypted Metadata i , j = A E S 128 _ C T R ( A E M K i , R P I i , j , Metadata ) {\displaystyle {\text{Associated Encrypted Metadata}}_{i,j}=AES128\_CTR(AEMK_{i},RPI_{i,j},{\text{Metadata}})} , where A E S 128 _ C T R ( Key, IV, Data ) {\displaystyle AES128\_CTR({\text{Key, IV, Data}})} denotes AES encryption with a 128-bit key in CTR mode. The Rolling Proximity Identifier and the Associated Encrypted Metadata are then combined and broadcast using BLE. Clients exchange and log these payloads. Once a registered health authority has confirmed the infection of a user, the user's Temporary Exposure Keys t e k i {\displaystyle tek_{i}} and their respective interval numbers i {\displaystyle i} for the past 14 days are uploaded to the central reporting server. Clients then download this report and individually recalculate every Rolling Proximity Identifier starting from interval number i {\displaystyle i} ,
Read more →
Sasha Luccioni

Alexandra Sasha Luccioni (née Vorobyova; born 1990) is a computer scientist specializing in the intersection of artificial intelligence (AI) and climate change. Her work focuses on quantifying the environmental impact of AI technologies and promoting sustainable practices in machine learning development. == Early life and education == Alexandra Sasha Vorobyova was born in the Ukrainian Soviet Socialist Republic in 1990. When she was four years old, her family relocated to Ontario, Canada. Her interest in science is influenced by her family's history; her mother, grandmother, and great-grandmother all pursued careers in scientific fields. Luccioni earned a B.A. in language science from University of Paris III: Sorbonne Nouvelle in 2010. Subsequently, she completed a M.S. in cognitive science, with a minor in natural language processing, at École normale supérieure in Paris in 2012. Luccioni obtained her PhD in cognitive computing from Université du Québec à Montréal (UQAM) in 2018. == Career == Luccioni began her professional career at Nuance Communications in 2017, where she focused on natural language processing (NLP) and machine learning (ML) techniques to enhance conversational agents. She then joined Morgan Stanley’s AI/ML Center of Excellence in 2018, working on explainable artificial intelligence (AI) and decision-making systems. In 2019, she became a postdoctoral researcher at Université de Montréal and Mila, collaborating with computer scientist Yoshua Bengio on a project titled This Climate Does Not Exist. This initiative used generative adversarial networks to visualize the effects of climate change. During this time, she also contributed to integrating fairness and accountability into machine learning education at Mila. Luccioni briefly worked with the United Nations Global Pulse in 2021, developing tools to monitor COVID-19 misinformation. Later that year, she joined Hugging Face as a research scientist. Her role includes quantifying the carbon footprint of AI systems, co-chairing the carbon working group in the Big Science project, and advancing responsible machine learning practices. She helped create "CodeCarbon," an open-source software tool that estimates the carbon emissions produced during the training and operation of machine learning models. In addition to her research, she has developed tools to measure the environmental impact of AI models, communicated findings through media engagements, and presented at international conferences, including a TED Talk. In 2024, she was listed on BBC 100 Women and Time 100 AI.
Read more →
François Chollet

François Chollet (French: [fʁɑ̃swa ʃoˈlɛ]; born 20 October 1989) is a French software engineer, artificial intelligence (AI) researcher, and former Senior Staff Engineer at Google. Chollet is the creator of the Keras deep-learning library released in 2015. His research focuses on computer vision, the application of machine learning to formal reasoning, abstraction, and how to achieve greater generality in artificial intelligence (AGI). == Education and career == In 2012, Chollet graduated with a Diplôme d'Ingénieur (Master of Engineering) from ENSTA Paris, a school of the Polytechnic Institute of Paris. In 2015, Chollet started working at Google shortly after releasing Keras. In 2019, he published the Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI) benchmark, which measures the ability of AI systems to solve novel reasoning problems. In 2024, Chollet launched ARC Prize, a US$1 million competition to solve the ARC-AGI benchmark. He left Google in November 2024 after more than 9 years with the company to found with Zapier co-founder Mike Knoop a new startup focused on developing AGI with program synthesis. In early 2025, Chollet announced the expansion of ARC Prize into a full-fledged non-profit foundation, to further the mission of guiding and accelerating research progress towards artificial general intelligence. == Books and publications == Chollet's research papers in artificial intelligence have been published at major conferences in the field, including the Conference on Computer Vision and Pattern Recognition (CVPR), the Conference on Neural Information Processing Systems (NeurIPS), and the International Conference on Learning Representations (ICLR). Chollet is the author of Xception: Deep Learning with Depthwise Separable Convolutions, which is among the top ten most cited papers in CVPR proceedings at more than 18,000 citations. Chollet is the author of the book Deep Learning with Python, which sold over 100,000 copies, and the co-author with Tomasz Kalinowski of Deep Learning With R. == Awards == On December 1, 2021, Chollet won the Global Swiss AI Award for breakthroughs in AI. In September 2024, Chollet was named by TIME as one of the 100 most influential people in AI.
Read more →
Restricted Boltzmann machine

A restricted Boltzmann machine (RBM) (also called a restricted Sherrington–Kirkpatrick model with external field or restricted stochastic Ising–Lenz–Little model) is a generative stochastic artificial neural network that can learn a probability distribution over its set of inputs. RBMs were initially proposed under the name Harmonium by Paul Smolensky in 1986, and rose to prominence after Geoffrey Hinton and collaborators used fast learning algorithms for them in the mid-2000s. RBMs have found applications in dimensionality reduction, classification, collaborative filtering, feature learning, topic modelling, immunology, and even many‑body quantum mechanics. They can be trained in either supervised or unsupervised ways, depending on the task. As their name implies, RBMs are a variant of Boltzmann machines, with the restriction that their neurons must form a bipartite graph: a pair of nodes from each of the two groups of units (commonly referred to as the "visible" and "hidden" units respectively) may have a symmetric connection between them; and there are no connections between nodes within a group. By contrast, "unrestricted" Boltzmann machines may have connections between hidden units. This restriction allows for more efficient training algorithms than are available for the general class of Boltzmann machines, in particular the gradient-based contrastive divergence algorithm. Restricted Boltzmann machines can also be used in deep learning networks. In particular, deep belief networks can be formed by "stacking" RBMs and optionally fine-tuning the resulting deep network with gradient descent and backpropagation. == Structure == The standard type of RBM has binary-valued (Boolean) hidden and visible units, and consists of a matrix of weights W {\displaystyle W} of size m × n {\displaystyle m\times n} . Each weight element ( w i , j ) {\displaystyle (w_{i,j})} of the matrix is associated with the connection between the visible (input) unit v i {\displaystyle v_{i}} and the hidden unit h j {\displaystyle h_{j}} . In addition, there are bias weights (offsets) a i {\displaystyle a_{i}} for v i {\displaystyle v_{i}} and b j {\displaystyle b_{j}} for h j {\displaystyle h_{j}} . Given the weights and biases, the energy of a configuration (pair of Boolean vectors) (v,h) is defined as E ( v , h ) = − ∑ i a i v i − ∑ j b j h j − ∑ i ∑ j v i w i , j h j {\displaystyle E(v,h)=-\sum _{i}a_{i}v_{i}-\sum _{j}b_{j}h_{j}-\sum _{i}\sum _{j}v_{i}w_{i,j}h_{j}} or, in matrix notation, E ( v , h ) = − a T v − b T h − v T W h . {\displaystyle E(v,h)=-a^{\mathrm {T} }v-b^{\mathrm {T} }h-v^{\mathrm {T} }Wh.} This energy function is analogous to that of a Hopfield network. As with general Boltzmann machines, the joint probability distribution for the visible and hidden vectors is defined in terms of the energy function as follows, P ( v , h ) = 1 Z e − E ( v , h ) {\displaystyle P(v,h)={\frac {1}{Z}}e^{-E(v,h)}} where Z {\displaystyle Z} is a partition function defined as the sum of e − E ( v , h ) {\displaystyle e^{-E(v,h)}} over all possible configurations, which can be interpreted as a normalizing constant to ensure that the probabilities sum to 1. The marginal probability of a visible vector is the sum of P ( v , h ) {\displaystyle P(v,h)} over all possible hidden layer configurations, P ( v ) = 1 Z ∑ { h } e − E ( v , h ) {\displaystyle P(v)={\frac {1}{Z}}\sum _{\{h\}}e^{-E(v,h)}} , and vice versa. Since the underlying graph structure of the RBM is bipartite (meaning there are no intra-layer connections), the hidden unit activations are mutually independent given the visible unit activations. Conversely, the visible unit activations are mutually independent given the hidden unit activations. That is, for m visible units and n hidden units, the conditional probability of a configuration of the visible units v, given a configuration of the hidden units h, is P ( v | h ) = ∏ i = 1 m P ( v i | h ) {\displaystyle P(v|h)=\prod _{i=1}^{m}P(v_{i}|h)} . Conversely, the conditional probability of h given v is P ( h | v ) = ∏ j = 1 n P ( h j | v ) {\displaystyle P(h|v)=\prod _{j=1}^{n}P(h_{j}|v)} . The individual activation probabilities are given by P ( h j = 1 | v ) = σ ( b j + ∑ i = 1 m w i , j v i ) {\displaystyle P(h_{j}=1|v)=\sigma \left(b_{j}+\sum _{i=1}^{m}w_{i,j}v_{i}\right)} and P ( v i = 1 | h ) = σ ( a i + ∑ j = 1 n w i , j h j ) {\displaystyle \,P(v_{i}=1|h)=\sigma \left(a_{i}+\sum _{j=1}^{n}w_{i,j}h_{j}\right)} where σ {\displaystyle \sigma } denotes the logistic sigmoid. The visible units of Restricted Boltzmann Machine can be multinomial, although the hidden units are Bernoulli. In this case, the logistic function for visible units is replaced by the softmax function P ( v i k = 1 | h ) = exp ⁡ ( a i k + Σ j W i j k h j ) Σ k ′ = 1 K exp ⁡ ( a i k ′ + Σ j W i j k ′ h j ) {\displaystyle P(v_{i}^{k}=1|h)={\frac {\exp(a_{i}^{k}+\Sigma _{j}W_{ij}^{k}h_{j})}{\Sigma _{k'=1}^{K}\exp(a_{i}^{k'}+\Sigma _{j}W_{ij}^{k'}h_{j})}}} where K is the number of discrete values that the visible values have. They are applied in topic modeling, and recommender systems. === Relation to other models === Restricted Boltzmann machines are a special case of Boltzmann machines and Markov random fields. The graphical model of RBMs corresponds to that of factor analysis. == Training algorithm == Restricted Boltzmann machines are trained to maximize the product of probabilities assigned to some training set V {\displaystyle V} (a matrix, each row of which is treated as a visible vector v {\displaystyle v} ), arg ⁡ max W ∏ v ∈ V P ( v ) {\displaystyle \arg \max _{W}\prod _{v\in V}P(v)} or equivalently, to maximize the expected log probability of a training sample v {\displaystyle v} selected randomly from V {\displaystyle V} : arg ⁡ max W E [ log ⁡ P ( v ) ] {\displaystyle \arg \max _{W}\mathbb {E} \left[\log P(v)\right]} The algorithm most often used to train RBMs, that is, to optimize the weight matrix W {\displaystyle W} , is the contrastive divergence (CD) algorithm due to Hinton, originally developed to train PoE (product of experts) models. The algorithm performs Gibbs sampling and is used inside a gradient descent procedure (similar to the way backpropagation is used inside such a procedure when training feedforward neural nets) to compute weight update. The basic, single-step contrastive divergence (CD-1) procedure for a single sample can be summarized as follows: Take a training sample v, compute the probabilities of the hidden units and sample a hidden activation vector h from this probability distribution. Compute the outer product of v and h and call this the positive gradient. From h, sample a reconstruction v' of the visible units, then resample the hidden activations h' from this. (Gibbs sampling step) Compute the outer product of v' and h' and call this the negative gradient. Let the update to the weight matrix W {\displaystyle W} be the positive gradient minus the negative gradient, times some learning rate: Δ W = ϵ ( v h T − v ′ h ′ T ) {\displaystyle \Delta W=\epsilon (vh^{\mathsf {T}}-v'h'^{\mathsf {T}})} . Update the biases a and b analogously: Δ a = ϵ ( v − v ′ ) {\displaystyle \Delta a=\epsilon (v-v')} , Δ b = ϵ ( h − h ′ ) {\displaystyle \Delta b=\epsilon (h-h')} . A Practical Guide to Training RBMs written by Hinton can be found on his homepage. == Stacked Restricted Boltzmann Machine == The difference between the Stacked Restricted Boltzmann Machines and RBM is that RBM has lateral connections within a layer that are prohibited to make analysis tractable. On the other hand, the Stacked Boltzmann consists of a combination of an unsupervised three-layer network with symmetric weights and a supervised fine-tuned top layer for recognizing three classes. The usage of Stacked Boltzmann is to understand Natural languages, retrieve documents, image generation, and classification. These functions are trained with unsupervised pre-training and/or supervised fine-tuning. Unlike the undirected symmetric top layer, with a two-way unsymmetric layer for connection for RBM. The restricted Boltzmann's connection is three-layers with asymmetric weights, and two networks are combined into one. Stacked Boltzmann does share similarities with RBM, the neuron for Stacked Boltzmann is a stochastic binary Hopfield neuron, which is the same as the Restricted Boltzmann Machine. The energy from both Restricted Boltzmann and RBM is given by Gibb's probability measure: E = − 1 2 ∑ i , j w i j s i s j + ∑ i θ i s i {\displaystyle E=-{\frac {1}{2}}\sum _{i,j}{w_{ij}{s_{i}}{s_{j}}}+\sum _{i}{\theta _{i}}{s_{i}}} . The training process of Restricted Boltzmann is similar to RBM. Restricted Boltzmann train one layer at a time and approximate equilibrium state with a 3-segment pass, not performing back propagation. Restricted Boltzmann uses both supervised and unsupervised on different RBM for pre-training for classification and recognition. The training uses contrastive divergence with
Read more →
Machine unlearning

Machine unlearning is a branch of machine learning focused on removing specific undesired element, such as private data, wrong or manipulated training data, outdated information, copyrighted material, harmful content, dangerous abilities, or misinformation, without needing to rebuild models from the ground up. Large language models, like the ones powering ChatGPT, may be asked not just to remove specific elements but also to unlearn a "concept," "fact," or "knowledge," which aren't easily linked to specific examples. New terms such as "model editing," "concept editing," and "knowledge unlearning" have emerged to describe this process. == History == Early research efforts were largely motivated by Article 17 of the GDPR, the European Union's privacy regulation commonly known as the "right to be forgotten" (RTBF), introduced in 2014. The GDPR did not anticipate that the development of large language models would make data erasure a complex task. This issue has since led to research on "machine unlearning," with a growing focus on removing copyrighted material, harmful content, dangerous capabilities, and misinformation. Just as early experiences in humans shape later ones, some concepts are more fundamental and harder to unlearn. A piece of knowledge may be so deeply embedded in the model's knowledge graph that unlearning it could cause internal contradictions, requiring adjustments to other parts of the graph to resolve them. Researchers have now also started studying unlearning in the context of removing incorrect or adversarially manipulated training data such as systematically biased labels or poisoning attacks. == Motivations == At present, machine unlearning is motivated by a growing range of concerns that extend well beyond the field's original focus on data privacy. A widely used taxonomy in the literature distinguishes two high-level categories of motivation. Access revocation covers cases where a data subject or rights holder requests the removal of data they own or control. This is most commonly associated with RTBF established by the European Union's General Data Protection Regulation (GDPR) and analogous legislation such as the California Consumer Privacy Act (CCPA). These regulations grant individuals the legal right to request erasure of their personal data from any system that has processed it, including models that were trained on it. Access revocation also encompasses the removal of copyrighted or pay-walled content that was incorporated into training corpora without the necessary licenses, a concern that has become prominent with the widespread use of largely web-scraped pre-training datasets. Model correction covers cases where the model exhibits undesirable behavior arising from the training data, regardless of any individual's request. This includes: Removal of toxic, biased, or unsafe outputs introduced by harmful content in the training set Correction of stale or factually incorrect associations, such as outdated knowledge encoded in a deployed model Removal of dangerous capabilities, such as detailed knowledge of the synthesis of chemical or biological agents Correction of the influence of data poisoning or adversarial attacks that have corrupted model behavior This second category has been formalized as corrective machine unlearning, which frames unlearning as a post-training mechanism for repairing the effects of bad or harmful training data. It is closely related to the AI safety literature, where data filtering alone has been found insufficient to prevent hazardous knowledge from being encoded in model weights, motivating unlearning as a complementary risk mitigation strategy. A further distinction has been drawn in the literature between removal {eliminating the influence of specific training data on model parameters) and suppression (preventing the model from generating specific outputs regardless of how that knowledge is encoded). These two goals are not equivalent: removing training data does not guarantee meaningful output suppression, and suppressing outputs does not constitute removal of the underlying training data's influence. == SISA Training == SISA is a training strategy consisting of four mechanisms designed to make machine unlearning more efficient by structuring how models are trained and updated. Its goal is to allow a system to remove the influence of specific data points without retraining an entire model from scratch. By reorganizing training data and workflows, SISA reduces the computational burden of unlearning requests. Sharding divides the training dataset into multiple disjoint subsets, or shards. Each shard is used to train a separate model instance. This ensures that a single data point affects only one shard, so unlearning it requires updating only the corresponding shard rather than the full model. Isolation refers to training each shard independently, with nothing shared across shards during the training process. This separation prevents cross-contamination between shards, ensuring that forgetting data in one shard does not require adjustments to any others. Slicing breaks the data within each shard into sequential slices and stores model states after each slice is trained on. When an unlearning request targets a piece of data, the system can roll back to the checkpoint before the point was seen and retrain only from that slice forward. This reduces retraining time even within a shard. Aggregation occurs at inference, when the model is queried. It combines the outputs of each shard to determine the output of the overall model. This is often through majority voting or averaging. This allows SISA-trained systems to behave like a single model despite being composed of multiple shard-level models. Together, these mechanisms enable machine learning systems to forget specific data points with far lower computational cost than full retraining. The trade-off is that sharding and slicing can lead to reduced model accuracy, worse generalization, and increased storage requirements for the intermediate checkpoints. This can be tolerable based on the needs of the individual or organization to comply with "right to be forgotten" or efficiently recover from backdoor attacks. == Algorithms == Machine unlearning algorithms are broadly categorized into exact and approximate methods, reflecting a fundamental trade-off between formal guarantees and computational tractability. === Exact Unlearning === Exact unlearning methods produce a model that is statistically indistinguishable from one retrained from scratch on the dataset with the forget data removed. The canonical framework for exact unlearning is SISA Training (Sharded, Isolated, Sliced, and Aggregated), introduced by Bourtoule et al. (2021). SISA partitions the training dataset into disjoint shards and trains a separate sub-model on each. At inference time, predictions are aggregated across sub-models. When an unlearning request is received, only the sub-model corresponding to the shard containing the target data requires retraining, reducing computational overhead proportionally to the number of shards. Exact methods provide the strongest guarantees but become prohibitively expensive for large pre-trained neural networks and are generally limited to settings where training can be structured in advance. === Approximate Unlearning === Approximate unlearning methods seek to produce a model whose behavior is sufficiently close to an exactly unlearned model without the cost of full retraining. These methods dominate practical applications. Common approaches include: Gradient Ascent: The model is fine-tuned by maximizing the loss on the forget set, directly degrading its performance on targeted data. This is the most direct approach but risks destabilizing performance on retained data. Random Labelling: The model is fine-tuned on the forget set using randomly shuffled labels, confusing its associations with the targeted data while producing a less aggressive weight shift than pure gradient ascent. Gradient Difference: Combines gradient ascent on the forget set with simultaneous gradient descent on the retain set, using the retain objective as a regularizer to preserve general model utility. KL Divergence Regularization: Minimizes the KL divergence between the outputs of the unlearned model and the original model on the retain set, anchoring behavior on data the model should remember. Weight Pruning and Fine-tuning: Parameters with the smallest L1-norm are pruned — targeting weights most weakly associated with general knowledge and potentially most associated with the forget set — followed by fine-tuning on the retain set to restore utility. Layer Reset and Fine-tuning: The first or last k layers are re-initialized to random weights and the model is subsequently fine-tuned on the retain set. This is a coarse but computationally simple approach. Selective Synaptic Dampening: Uses influence functions to estimate the effect of individual trainin
Read more →
Top 10 AI Presentation Makers Compared (2026)

Trying to pick the best AI presentation maker? An AI presentation maker is software that uses machine learning to help you get more done — it scales effortlessly from a single task to thousands. The best picks balance beginner-friendly simplicity with the depth power users need, and they ship updates often. Whether you are a beginner or a pro, the right AI presentation maker slots into your workflow and pays for itself fast. Read on for hands-on impressions, pricing tiers, and the standout features that matter.
Read more →
Levenshtein automaton

In computer science, a Levenshtein automaton for a string w and a number n is a finite-state automaton that can recognize the set of all strings whose Levenshtein distance from w is at most n. That is, a string x is in the formal language recognized by the Levenshtein automaton if and only if x can be transformed into w by at most n single-character insertions, deletions, and substitutions. == Applications == Levenshtein automata may be used for spelling correction, by finding words in a given dictionary that are close to a misspelled word. In this application, once a word is identified as being misspelled, its Levenshtein automaton may be constructed, and then applied to all of the words in the dictionary to determine which ones are close to the misspelled word. If the dictionary is stored in compressed form as a trie, the time for this algorithm (after the automaton has been constructed) is proportional to the number of nodes in the trie, significantly faster than using dynamic programming to compute the Levenshtein distance separately for each dictionary word. It is also possible to find words in a regular language, rather than a finite dictionary, that are close to a given target word, by computing the Levenshtein automaton for the word, and then using a Cartesian product construction to combine it with an automaton for the regular language, giving an automaton for the intersection language. Alternatively, rather than using the product construction, both the Levenshtein automaton and the automaton for the given regular language may be traversed simultaneously using a backtracking algorithm. Levenshtein automata are used in Lucene for full-text searches that can return relevant documents even if the query is misspelled. == Construction == For any fixed constant n, the Levenshtein automaton for w and n may be constructed in time O(|w|). Mitankin studies a variant of this construction called the universal Levenshtein automaton, determined only by a numeric parameter n, that can recognize pairs of words (encoded in a certain way by bitvectors) that are within Levenshtein distance n of each other. Touzet proposed an effective algorithm to build this automaton. Yet a third finite automaton construction of Levenshtein (or Damerau–Levenshtein) distance are the Levenshtein transducers of Hassan et al., who show finite state transducers implementing edit distance one, then compose these to implement edit distances up to some constant.
Read more →
AI Pair Programmers: Free vs Paid (2026)

Trying to pick the best AI pair programmer? An AI pair programmer is software that uses machine learning to help you get more done — it scales effortlessly from a single task to thousands. The best picks balance beginner-friendly simplicity with the depth power users need, and they ship updates often. Whether you are a beginner or a pro, the right AI pair programmer slots into your workflow and pays for itself fast. This guide breaks down the top picks, their pros and cons, and who each one is best for.
Read more →
Tail latency

Tail latency is a term used to describe the high-percentile response times seen in a system. This is usually measured at the 95th, 99th, or 99.9th percentile, not the average latency. In distributed systems, cloud computing, and large-scale web services, even a small number of slow requests can make the user experience and system performance much worse. Tail latency often happens because of things like resource contention, network variability, garbage collection pauses, and hardware heterogeneity. A major problem in system design is managing tail latency, because lowering average latency doesn't always make the worst-case performance better. To lessen its effects, people often use techniques like request hedging, replication, load balancing, and adaptive timeouts. In latency-sensitive applications like search engines, financial systems, and real-time services, where service-level objectives (SLOs) are often based on high-percentile latencies, it is especially important to understand and improve tail latency.
Read more →
Top 10 AI Background Removers Compared (2026)

Curious about the best AI background remover? An AI background remover is software that uses machine learning to help you get more done — it combines speed, accuracy, and an interface that just works. Hands-on testing shows real-world results vary, so a short free trial is the smartest way to decide. Whether you are a beginner or a pro, the right AI background remover slots into your workflow and pays for itself fast. This guide breaks down the top picks, their pros and cons, and who each one is best for.
Read more →
DeepL Translator

DeepL is a German AI research company known for its language AI platform, which includes DeepL Translator and DeepL Voice, and for DeepL Agent, an AI agent capable of planning workflows and using office systems and tools autonomously, in response to natural language instructions. Its algorithm uses the transformer architecture. It offers a paid subscription for additional features and access to its translation application programming interface. DeepL was founded in 2017 by Jaroslaw Kutylowski and is a unicorn, valued at $2 billion after a Series C funding round raised $300 million in May 2024. Its more than 200,000 business customers include a large proportion of the Fortune 500. == History == The translating system was first developed within Linguee by a team led by Chief Technology Officer Jarosław Kutyłowski in 2016. It was launched as DeepL Translator on 28 August 2017 and offered translations between English, German, French, Spanish, Italian, Polish and Dutch. At its launch, it claimed to have surpassed its competitors in blind tests and BLEU scores, including Google Translate, Amazon Translate, Microsoft Translator and Facebook's translation feature. With the release of DeepL in 2017, Linguee's company name was changed to DeepL GmbH, and it is also financed by advertising on its sister site, linguee.com. Support for Portuguese and Russian was added on 5 December 2018. In July 2019, Jarosław Kutyłowski became the CEO of DeepL GmbH and restructured the company into a Societas Europaea in 2021. Translation software for Microsoft Windows and macOS was released in September 2019. Support for Chinese (simplified) and Japanese was added on 19 March 2020, which the company claimed to have surpassed the aforementioned competitors as well as Baidu and Youdao. Then, 13 more European languages were added in March 2021: Bulgarian, Czech, Danish, Estonian, Finnish, Greek, Hungarian, Latvian, Lithuanian, Romanian, Slovak, Slovenian, and Swedish, bringing the total number of supported languages to 24. On 25 May 2022, support for Indonesian and Turkish was added, and support for Ukrainian was added on 14 September 2022. In January 2023, the company reached a valuation of 1 billion euro and became the most valued startup company in Cologne. At the end of the month, support for Korean and Norwegian (Bokmål) was also added. In May 2024, the company announced an investment of US$300 million at AI. In January 2026, more languages were supported, including Luxembourgish and Irish. == Services == === Translation method === The service uses a proprietary algorithm with convolutional neural networks (CNNs) that have been trained with the Linguee database. According to the developers, the service uses a newer improved architecture of neural networks, resulting in a more natural sound of translations than by competing services. The translation is generated using a supercomputer that reaches 5.1 petaflops and is operated in Iceland with hydropower. DeepL's data centers are located at the EcoDataCenter in Falun, Sweden, which is a data center for sustainability. In general, CNNs are slightly more suitable for long coherent word sequences, but they have so far not been used by the competition because of their weaknesses compared to recurrent neural networks. The weaknesses of DeepL are compensated for by supplemental techniques, some of which are publicly known. === Translator and subscription === The translator can be used for free with a maximum limit of 1,500 characters per translation. Microsoft Word and PowerPoint files in Office Open XML file formats (.docx and .pptx) and PDF files up to 5MB in size can also be translated. It offers paid subscription DeepL Pro, which has been available since March 2018 and includes application programming interface access and a software plug-in for computer-assisted translation tools, including SDL Trados Studio. Unlike the free version, translated texts are stated to not be saved on the server; also, the character limit is removed. The monthly pricing model includes a set amount of text, with texts beyond that being calculated according to the number of characters. ==== Supported languages ==== As of May 2026, the translation service supports the following languages: Additionally, these languages are currently in beta, indicated by an asterisk after their name in the language picker: === DeepL Write === In November 2022, DeepL launched a tool to improve monolingual texts in English and German, called DeepL Write. In December, the company removed access and informed journalists that it was only for internal use and that DeepL Write would be relaunched in early 2023. The public beta version was then released on January 17, 2023. In the summer of 2024, DeepL announced the availability of two more languages in DeepL Write: French and Spanish. By January 2024, DeepL had added an additional two: Portuguese (European and Brazilian) and Italian. === DeepL Agent === In November 2025, DeepL launched an AI agent called DeepL Agent which is capable of operating business applications in a human-like manner. == Reception == The reception of DeepL has been generally positive. TechCrunch appreciates it for the accuracy of its translations and stating that it was more accurate and nuanced than Google Translate. Le Monde thanks its developers for translating French text into more "French-sounding" expressions. RTL Z stated that DeepL Translator "offers better translations […] when it comes to Dutch to English and vice versa". La Repubblica, and a Latin American website, "WWWhat's new?", showed praise as well. A 2018 paper by the University of Bologna evaluated the Italian-to-German translation capabilities and found the preliminary results to be similar in quality to Google Translate. In September 2021, Slator remarked that the language industry response was more measured than the press and noted that DeepL is still highly regarded by users. A reviewer noted in 2018 that DeepL had far fewer languages available for translation than competing products. == Awards and honors == DeepL won the 2020 Webby Award for Best Practices and the 2020 Webby Award for Technical Achievement (Apps, Mobile, and Features), both in the category Apps, Mobile & Voice. In April 2025, DeepL was featured in the Forbes AI 50 list.
Read more →
Probabilistic context-free grammar

In theoretical linguistics and computational linguistics, probabilistic context free grammars (PCFGs) extend context-free grammars, similar to how hidden Markov models extend regular grammars. Each production is assigned a probability. The probability of a derivation (parse) is the product of the probabilities of the productions used in that derivation. These probabilities can be viewed as parameters of the model, and for large problems it is convenient to learn these parameters via machine learning. A probabilistic grammar's validity is constrained by context of its training dataset. PCFGs originated from grammar theory, and have application in areas as diverse as natural language processing to the study the structure of RNA molecules and design of programming languages. Designing efficient PCFGs has to weigh factors of scalability and generality. Issues such as grammar ambiguity must be resolved. The grammar design affects results accuracy. Grammar parsing algorithms have various time and memory requirements. == Definitions == Derivation: The process of recursive generation of strings from a grammar. Parsing: Finding a valid derivation using an automaton. Parse Tree: The alignment of the grammar to a sequence. An example of a parser for PCFG grammars is the pushdown automaton. The algorithm parses grammar nonterminals from left to right in a stack-like manner. This brute-force approach is not very efficient. In RNA secondary structure prediction variants of the Cocke–Younger–Kasami (CYK) algorithm provide more efficient alternatives to grammar parsing than pushdown automata. Another example of a PCFG parser is the Stanford Statistical Parser which has been trained using Treebank. == Formal definition == Similar to a CFG, a probabilistic context-free grammar G can be defined by a quintuple: G = ( M , T , R , S , P ) {\displaystyle G=(M,T,R,S,P)} where M is the set of non-terminal symbols T is the set of terminal symbols R is the set of production rules S is the start symbol P is the set of probabilities on production rules == Relation with hidden Markov models == PCFGs models extend context-free grammars the same way as hidden Markov models extend regular grammars. The Inside-Outside algorithm is an analogue of the Forward-Backward algorithm. It computes the total probability of all derivations that are consistent with a given sequence, based on some PCFG. This is equivalent to the probability of the PCFG generating the sequence, and is intuitively a measure of how consistent the sequence is with the given grammar. The Inside-Outside algorithm is used in model parametrization to estimate prior frequencies observed from training sequences in the case of RNAs. Dynamic programming variants of the CYK algorithm find the Viterbi parse of a RNA sequence for a PCFG model. This parse is the most likely derivation of the sequence by the given PCFG. == Grammar construction == Context-free grammars are represented as a set of rules inspired from attempts to model natural languages. The rules are absolute and have a typical syntax representation known as Backus–Naur form. The production rules consist of terminal { a , b } {\displaystyle \left\{a,b\right\}} and non-terminal S symbols and a blank ϵ {\displaystyle \epsilon } may also be used as an end point. In the production rules of CFG and PCFG the left side has only one nonterminal whereas the right side can be any string of terminal or nonterminals. In PCFG nulls are excluded. An example of a grammar: S → a S , S → b S , S → ϵ {\displaystyle S\to aS,S\to bS,S\to \epsilon } This grammar can be shortened using the '|' ('or') character into: S → a S | b S | ϵ {\displaystyle S\to aS|bS|\epsilon } Terminals in a grammar are words and through the grammar rules a non-terminal symbol is transformed into a string of either terminals and/or non-terminals. The above grammar is read as "beginning from a non-terminal S the emission can generate either a or b or ϵ {\displaystyle \epsilon } ". Its derivation is: S ⇒ a S ⇒ a b S ⇒ a b b S ⇒ a b b {\displaystyle S\Rightarrow aS\Rightarrow abS\Rightarrow abbS\Rightarrow abb} Ambiguous grammar may result in ambiguous parsing if applied on homographs since the same word sequence can have more than one interpretation. Pun sentences such as the newspaper headline "Iraqi Head Seeks Arms" are an example of ambiguous parses. One strategy of dealing with ambiguous parses (originating with grammarians as early as Pāṇini) is to add yet more rules, or prioritize them so that one rule takes precedence over others. This, however, has the drawback of proliferating the rules, often to the point where they become difficult to manage. Another difficulty is overgeneration, where unlicensed structures are also generated. Probabilistic grammars circumvent these problems by ranking various productions on frequency weights, resulting in a "most likely" (winner-take-all) interpretation. As usage patterns are altered in diachronic shifts, these probabilistic rules can be re-learned, thus updating the grammar. Assigning probability to production rules makes a PCFG. These probabilities are informed by observing distributions on a training set of similar composition to the language to be modeled. On most samples of broad language, probabilistic grammars where probabilities are estimated from data typically outperform hand-crafted grammars. CFGs when contrasted with PCFGs are not applicable to RNA structure prediction because while they incorporate sequence-structure relationship they lack the scoring metrics that reveal a sequence structural potential == Weighted context-free grammar == A weighted context-free grammar (WCFG) is a more general category of context-free grammar, where each production has a numeric weight associated with it. The weight of a specific parse tree in a WCFG is the product (or sum ) of all rule weights in the tree. Each rule weight is included as often as the rule is used in the tree. A special case of WCFGs are PCFGs, where the weights are (logarithms of ) probabilities. An extended version of the CYK algorithm can be used to find the "lightest" (least-weight) derivation of a string given some WCFG. When the tree weight is the product of the rule weights, WCFGs and PCFGs can express the same set of probability distributions. == Applications == === RNA structure prediction === Since the 1990s, PCFG has been applied to model RNA structures. Energy minimization and PCFG provide ways of predicting RNA secondary structure with comparable performance. However structure prediction by PCFGs is scored probabilistically rather than by minimum free energy calculation. PCFG model parameters are directly derived from frequencies of different features observed in databases of RNA structures rather than by experimental determination as is the case with energy minimization methods. The types of various structure that can be modeled by a PCFG include long range interactions, pairwise structure and other nested structures. However, pseudoknots can not be modeled. PCFGs extend CFG by assigning probabilities to each production rule. A maximum probability parse tree from the grammar implies a maximum probability structure. Since RNAs preserve their structures over their primary sequence, RNA structure prediction can be guided by combining evolutionary information from comparative sequence analysis with biophysical knowledge about a structure plausibility based on such probabilities. Also search results for structural homologs using PCFG rules are scored according to PCFG derivations probabilities. Therefore, building grammar to model the behavior of base-pairs and single-stranded regions starts with exploring features of structural multiple sequence alignment of related RNAs. S → a S a | b S b | a a | b b {\displaystyle S\to aSa|bSb|aa|bb} The above grammar generates a string in an outside-in fashion, that is the basepair on the furthest extremes of the terminal is derived first. So a string such as a a b a a b a a {\displaystyle aabaabaa} is derived by first generating the distal a's on both sides before moving inwards: S ⇒ a S a ⇒ a a S a a ⇒ a a b S b a a ⇒ a a b a a b a a {\displaystyle S\Rightarrow aSa\Rightarrow aaSaa\Rightarrow aabSbaa\Rightarrow aabaabaa} A PCFG model extendibility allows constraining structure prediction by incorporating expectations about different features of an RNA . Such expectation may reflect for example the propensity for assuming a certain structure by an RNA. However incorporation of too much information may increase PCFG space and memory complexity and it is desirable that a PCFG-based model be as simple as possible. Every possible string x a grammar generates is assigned a probability weight P ( x | θ ) {\displaystyle P(x|\theta )} given the PCFG model θ {\displaystyle \theta } . It follows that the sum of all probabilities to all possible grammar productions is ∑ x P ( x | θ ) = 1 {\displaystyle \sum _{\text{x}}P(x|\theta )=1} . The scores
Read more →
Rabbit r1

The Rabbit r1 is an artificial intelligence personal assistant device developed by the American technology startup Rabbit Inc and co-designed by Teenage Engineering. It was announced at the 2024 Consumer Electronics Show as a handheld device intended to perform digital tasks through voice commands, touch interaction, and web-based AI agents. The r1 was marketed around Rabbit's concept of a "large action model" (LAM), which the company described as software able to operate websites and services on behalf of users. The device runs rabbitOS, an operating system based on the Android Open Source Project. Its services have included AI search, image recognition, voice interaction, music playback, rideshare and food-ordering integrations, and later experimental web-agent features such as LAM Playground and teach mode. Initial reviews were largely negative, with reviewers criticizing the device's limited functionality, bugs, and unclear advantages over a smartphone. Critics also questioned Rabbit's claims after the r1 software was shown to run on an Android phone. Rabbit continued to issue software updates after launch, including rabbitOS 2 in September 2025, which introduced a redesigned card-based interface, gesture navigation, and a "creations" feature for generating small software tools and experiences on the device. Rabbit Inc was founded by Jesse Lyu Cheng. == Hardware == Display: A 2.88-inch touchscreen for interactive user input. Input: push-to-talk button to activate voice commands; scroll wheel; Gyroscope; Magnetometer; Accelerometer; GPS. Camera: 8 MP single camera, with a resolution of 3264x2448, allowing for the connected external AI to use computer vision. Audio: Equipped with a speaker and dual microphones for audio interaction. Connectivity: Supports Wi-Fi and cellular connections via a SIM card slot to access internet services. Processor: Runs on a 2.3GHz MediaTek Helio P35 processor. Memory: Contains 4GB of RAM for operational tasks. Storage: Offers 128GB of internal storage for data. Ports: Utilizes a USB-C port for charging and data connections. == Software == The Rabbit r1 runs rabbitOS, which is based on the Android Open Source Project (AOSP), specifically Android 13. Rabbit founder Jesse Lyu described rabbitOS as a "very bespoke AOSP" after reports that the r1's software could be run on a conventional Android phone. Rabbit described the r1 as using a large action model (LAM), a type of AI agent intended to perform tasks across software interfaces rather than only answer questions. At launch, the device supported a limited set of services, including AI search, vision features, music playback, and some third-party integrations. Perplexity.ai was one of the AI services used to answer user queries. In 2024, Rabbit released several software updates that added features and attempted to address early criticism of the device. In July 2024, the company launched "beta rabbit", an advanced search and conversation mode for more complex queries. In October 2024, it released LAM Playground, a web-based agent feature intended to let the r1 operate websites on behalf of users. Reviewers found the feature experimental; Android Authority reported that it could perform some navigation tasks but struggled with CAPTCHAs, loops, and unintended behavior. In November 2024, Rabbit introduced a beta "teach mode", which allowed users to demonstrate web-based tasks in the Rabbithole web portal and later ask the r1 to repeat them. The company described teach mode as experimental, and The Verge noted that Rabbit warned users that results could be unpredictable and that CAPTCHA-protected sites could cause problems. Rabbit released rabbitOS 2 in September 2025. The update redesigned the interface around a card-based layout, added additional touchscreen gestures, and introduced "creations", a feature that lets users generate simple software tools, games, and interfaces through natural-language prompts. Coverage of the update described it as a major software overhaul rather than new hardware. == Reception == === Funding === Rabbit raised $20 million in funding from Khosla Ventures, Synergis Capital and Kakao Investment in October 2023. The company announced an additional $10 million in funding in December 2023. === Sales === Following its announcement at the 2024 Consumer Electronics Show, 130,000 units were sold. On August 13, 2024, Rabbit announced that sales of r1 had expanded to the entire European Union (except Malta) and United Kingdom. On August 21, 2024, sales of r1 expanded to Singapore. === Reviews === The r1 was met with strong criticism immediately after Rabbit began shipping the device. Some reviews questioned what the device was able to do that a smartphone could not, while comparing it to the similar Humane Ai Pin. YouTuber Marques Brownlee called the device "barely reviewable". Android Authority's Mishaal Rahman managed to install Rabbit r1's software on a Pixel 6a smartphone, after a tipster shared an APK file. The Verge echoed the claims made by Rahman. In response, Lyu published statements confirming its use of Android, but denying that the r1 is an Android app. Mashable called its Vision features impressive, but said that "these praise-worthy features are overshadowed by buggy performance". Ars Technica wrote a blog post claiming "the company is blocking access from bootleg APKs". TechCrunch gave a slightly more positive review, calling the device a "fun peep at a possible future", but could not "advise anyone to buy one now." Shortly after the launch of r1, Rabbit began a weekly cadence of software updates to address much of the criticism from the early reviews, including "battery and GPS performance, time zone selection, and more". Digital Trends said the Magic Camera feature "takes the most mundane, ordinary, and badly composed photos and makes something fun and eye-catching from them." Mashable said the "beta rabbit" feature "makes Rabbit R1 more conversational and intelligent". Later coverage noted that Rabbit continued to update the r1 after its poorly received launch. The Verge reported in September 2024 that about 5,000 of roughly 100,000 purchasers were using the device at any given moment, citing Lyu, and described the product as having launched before it was ready. In 2025, coverage of rabbitOS 2 described the update as an attempt to reset the device's software experience after the criticism of its original release. == Controversies == === GAMA project === Rabbit Inc has garnered attention due to allegations surrounding its funding and the company's past projects. The company came under scrutiny when Stephen Findeisen, known as Coffeezilla on YouTube, published a video in May 2024, alleging that Rabbit Incorporation was "built on a scam". Rabbit Incorporation, initially named Cyber Manufacturing Co, rebranded just two months before launching the Rabbit R1. The company, under its former name, raised $6 million in November 2021 for a project called GAMA, described as a "Next Generation NFT Project." Jesse Lyu, the CEO of Rabbit Incorporation, referred to GAMA as a "fun little project." Coffeezilla, who investigates influencer scams, highlighted old Clubhouse recordings of Jesse Lyu discussing the GAMA project. In these recordings, Lyu emphasized the substantial funding behind GAMA and its potential to be a revolutionary, carbon-negative cryptocurrency. Coffeezilla questioned the whereabouts of the funds raised for GAMA, estimating that approximately $1 million in refunds to investors remained unresolved. He suggested that the rebranding to Rabbit Incorporation and the shift to developing the Rabbit R1 were attempts to divert from the GAMA project's issues. In response to Coffeezilla's inquiries, Rabbit Incorporation stated that the $6 million raised was used for the GAMA project. The company said that NFTs cannot be refunded unless the owner agrees to "burn" them on the blockchain. Rabbit Incorporation also said that the GAMA project was open-sourced and returned to the community, aligning with community feedback. They also mentioned that efforts to buy back NFTs were made to counteract malicious trading and maintain market stability. === Security === In June 2024, Engadget reported that the Rabbitude team, a community reverse engineering project, had gained access to the r1's codebase revealing that r1's software contained several hardcoded API keys in its code for ElevenLabs, Microsoft Azure, Yelp, and Google Maps, potentially allowing unauthorized access to r1 responses, including those containing the users' personal information. For a short time, Rabbit immediately began revoking and rotating those secrets and confirmed that the code was leaked by an employee who had "been terminated and remains under investigation". In July 2024, the company revealed that all user chats and device pairing data were logged on the r1 with no ability to delete them. This meant that lost or stolen devices could be used to extract user
Read more →
Margaret Mitchell (scientist)

Margaret Mitchell is a computer scientist who works on algorithmic bias and fairness in machine learning. She is most well known for her work on automatically removing undesired biases concerning demographic groups from machine learning models, as well as more transparent reporting of their intended use. == Education == Mitchell obtained a bachelor's degree in linguistics from Reed College, Portland, Oregon, in 2005. After having worked as a research assistant at the OGI School of Science and Engineering for two years, she subsequently obtained a Master's in Computational Linguistics from the University of Washington in 2009. She enrolled in a PhD program at the University of Aberdeen, where she wrote a doctoral thesis on the topic of Generating Reference to Visible Objects, graduating in 2013. == Career and research == Mitchell is best known for her work on fairness in machine learning and methods for mitigating algorithmic bias. This includes her work on introducing the concept of 'Model Cards' for more transparent model reporting, and methods for debiasing machine learning models using adversarial learning. Margaret Mitchell created the framework for recognizing and avoiding biases by testing with a variable for the group of interest, predictor and an adversary. In 2012, Mitchell joined the Human Language Technology Center of Excellence at Johns Hopkins University as a postdoctoral researcher, before taking up a position at Microsoft Research in 2013. At Microsoft, Mitchell was the research lead of the Seeing AI project, an app that offers support for the visually impaired by narrating texts and images. In November 2016, she became a senior research scientist at Google Research and Machine intelligence. While at Google, she founded and co-led the Ethical Artificial Intelligence team together with Timnit Gebru. In May 2018, she represented Google in the Partnership on AI. In February 2018, she gave a TED talk on "How we can build AI to help humans, not hurt us". In January 2021, after Timnit Gebru's termination from Google, Mitchell reportedly used a script to search through her corporate account and download emails that allegedly documented discriminatory incidents involving Gebru. An automated system locked Mitchell's account in response. In response to media attention Google claimed that she "exfiltrated thousands of files and shared them with multiple external accounts". After a five-week investigation, Mitchell was fired. Prior to her dismissal, Mitchell had been a vocal advocate for diversity at Google, and had voiced concerns about research censorship at the company. In late 2021, she joined AI start-up Hugging Face. Mitchell is a co-founder of Widening NLP, a special interest group within the Association for Computational Linguistics (ACL) seeking to increase the proportion of women and minorities working in natural language processing; and Computational Linguistics and Clinical Psychology, an annual workshop within the ACL that brings together clinicians and computational linguists to advance the state of the art in clinical psychology.
Read more →
How to Choose an AI Chatbot

Looking for the best AI chatbot? An AI chatbot is software that uses machine learning to help you get more done — it can save you hours every week by automating repetitive work. Most options offer a generous free tier, with paid plans unlocking higher limits, faster processing, and team features. Whether you are a beginner or a pro, the right AI chatbot slots into your workflow and pays for itself fast. Read on for hands-on impressions, pricing tiers, and the standout features that matter.
Read more →