Machine Learning | Techlist.io

metaNov 18, 2025Original article

Efficient Optimization With Ax, an Open Platform for Adaptive Experimentation (opens in new tab)

Meta has released Ax 1.0, an open-source platform designed to automate and optimize complex, resource-intensive experimentation through machine learning. By utilizing Bayesian optimization, the platform helps researchers navigate vast configuration spaces to improve AI models, infrastructure, and hardware design efficiently. The release aims to bridge the gap between sophisticated mathematical theory and the practical requirements of production-scale engineering. ## Real-World Experimentation and Utility * Ax is used extensively at Meta for diverse tasks, including tuning hyperparameter configurations, discovering optimal data mixtures for Generative AI, and optimizing compiler flags. * The platform is built to handle the logistical "overhead" of experimentation, such as managing experiment states, automating orchestration, and providing diagnostic tools. * It supports multi-objective optimization, allowing users to balance competing metrics and enforce "guardrail" constraints rather than just maximizing a single value. * Applications extend beyond software to physical engineering, such as optimizing design parameters for AR/VR hardware. ## System Insight and Analysis * Beyond finding optimal points, Ax serves as a diagnostic tool to help researchers understand the underlying behavior of their systems. * It includes built-in visualizations for Pareto frontiers, which illustrate the trade-offs between different metrics. * Sensitivity analysis tools identify which specific input parameters have the greatest impact on the final results. * The platform provides automated plots and tables to track optimization progress and visualize the effect of parameters across the entire input space. ## Technical Methodology and Architecture * Ax utilizes Bayesian optimization, an iterative approach that balances "exploration" (sampling new areas) with "exploitation" (refining known good areas). * The platform relies on **BoTorch** for its underlying Bayesian components and typically employs **Gaussian processes (GP)** as surrogate models. * GPs are preferred because they can make accurate predictions and quantify uncertainty even when provided with very few data points. * The system uses an **Expected Improvement (EI)** acquisition function to calculate the potential value of new configurations compared to the current best-known result. * This surrogate-based approach is designed to scale to high-dimensional settings involving hundreds of tunable parameters where traditional search methods are too costly. To begin implementing these methods, developers can install the platform via `pip install ax-platform`. Ax 1.0 provides a robust framework for moving cutting-edge optimization research directly into production environments.

Machine Learning AI Bayesian Optimization Adaptive Experimentation+5

googleNov 18, 2025Original article

Real-time speech-to-speech translation (opens in new tab)

Google DeepMind and Google Core ML have developed an innovative end-to-end speech-to-speech translation (S2ST) model that enables real-time, voice-preserved communication with only a two-second delay. By replacing traditional cascaded pipelines with a streaming architecture trained on time-synchronized data, the system overcomes long-standing issues of high latency and accumulated errors. This advancement represents a significant shift toward natural, fluid cross-language dialogue that retains the original speaker's personality. ## Limitations of Cascaded S2ST Traditional real-time translation systems typically rely on a cascaded chain of three distinct AI models: Automatic Speech Recognition (ASR), Automatic Speech Translation (AST), and Text-to-Speech (TTS). This approach suffers from several critical drawbacks: * **High Latency:** Processing through three separate stages results in a 4–5 second delay, forcing users into unnatural, turn-based interactions. * **Error Propagation:** Inaccuracies in the initial transcription or translation phase accumulate, often leading to garbled or incorrect final audio output. * **Loss of Identity:** General-purpose TTS engines generate generic voices, stripping the communication of the original speaker’s unique vocal characteristics. ## Time-Synced Data Acquisition Pipeline To train an end-to-end model capable of low-latency output, researchers created a scalable pipeline that transforms raw audio into a specialized time-synchronized dataset. * **Alignment Multi-mapping:** The process uses forced alignment algorithms to map source audio to source text, source text to translated text, and finally, translated text to generated speech. * **Voice Preservation:** A custom TTS engine generates the target language audio while intentionally preserving the vocal characteristics of the original speaker. * **Strict Validation:** Automated filters discard any segments where alignments fail or where the translated audio cannot meet specific real-time delay requirements. * **Data Augmentation:** The training set is further refined using techniques such as sample rate reduction, denoising, and reverberation to ensure the model performs well in real-world environments. ## End-to-End Streaming Architecture The model’s architecture is designed for continuous audio streams, leveraging the AudioLM framework and fundamental transformer blocks to make real-time decisions. * **Streaming Encoder:** This component summarizes source audio data by focusing on the preceding 10-second window of input. * **Streaming Decoder:** This module predicts translated audio autoregressively, utilizing compressed encoder states and previous predictions to maintain flow. * **RVQ Audio Tokens:** The system represents audio as a 2D set of Residual Vector Quantization (RVQ) tokens, where the X-axis represents time and the Y-axis represents audio quality/fidelity. * **SpectroStream Integration:** By using SpectroStream codec technology, the model manages hierarchical audio representations, allowing it to prioritize the sequential output of audio segments for immediate playback. This technology effectively bridges the gap between high-quality translation and real-time responsiveness. For developers and researchers in the field, the transition from modular cascaded systems to end-to-end streaming architectures—supported by rigorous time-aligned datasets—is the recommended path for achieving truly seamless human-to-human cross-language communication.

Machine Learning AI Transformer Text To Speech+5

googleNov 12, 2025Original article

Separating natural forests from other tree cover with AI for deforestation-free supply chains (opens in new tab)

Researchers from Google DeepMind and Google Research have developed "Natural Forests of the World 2020," an AI-powered global map that distinguishes natural ecosystems from commercial tree plantations. By utilizing high-resolution satellite data and machine learning, the project provides a critical 10-meter resolution baseline to support deforestation-free supply chain regulations like the EUDR. This tool enables governments and companies to monitor biodiversity-rich areas with unprecedented accuracy, ensuring that natural forests are protected from industrial degradation. **The Limitation of Traditional Tree Cover Maps** * Existing maps frequently conflate all woody vegetation into a generic "tree cover" category, leading to "apples-to-oranges" comparisons between different land types. * This lack of distinction makes it difficult to differentiate between the harvesting of short-term plantations and the permanent loss of ancient, biodiversity-rich natural forests. * Precise mapping is now a legal necessity due to regulations like the European Union Regulation on Deforestation-free Products (EUDR), which bans products from land deforested or degraded after December 31, 2020. **The MTSViT Modeling Approach** * To accurately identify forest types, researchers developed the Multi-modal Temporal-Spatial Vision Transformer (MTSViT). * Rather than relying on a single snapshot, the AI "observes" 1280 x 1280 meter patches over the course of a year to identify seasonal, spectral, and textural signatures. * The model integrates multi-modal data, including Sentinel-2 satellite imagery, topographical information (such as elevation and slope), and specific geographical coordinates. * This temporal-spatial analysis allows the AI to recognize the complex patterns of natural forests that distinguish them from the uniform, fast-growing structures of commercial plantations. **Dataset Scale and Global Validation** * The model was trained on a massive dataset comprising over 1.2 million global patches at 10-meter resolution. * The final map provides seamless global coverage, achieving a best-in-class validation accuracy of 92.2% against an independent global dataset. * The research was a collaborative effort involving the World Resources Institute and the International Institute for Applied Systems Analysis to ensure scientific rigor and practical utility. The "Natural Forests of the World 2020" dataset is publicly available via Google Earth Engine and other open repositories. Organizations should leverage this high-resolution baseline to conduct environmental due diligence, support government monitoring, and target conservation efforts in preparation for global climate milestones like COP30.

Machine Learning AI Computer Vision Remote Sensing+3

googleNov 11, 2025Original article

Differentially private machine learning at scale with JAX-Privacy (opens in new tab)

Google DeepMind and Google Research have announced the release of JAX-Privacy 1.0, a high-performance library designed to scale differentially private (DP) machine learning. By leveraging JAX’s native parallelization and functional programming model, the toolkit enables researchers to train large-scale foundation models while maintaining rigorous privacy guarantees. This version introduces modular components for advanced algorithms and empirical auditing, making private training both computationally efficient and verifiable across distributed environments. ### Scaling Differential Privacy with JAX * The library is built directly on the JAX ecosystem, integrating seamlessly with Flax for neural network architectures and Optax for optimization. * It utilizes JAX’s `vmap` for automatic vectorization and `shard_map` for single-program multiple-data (SPMD) parallelization, allowing DP primitives to scale across multiple accelerators. * By using just-in-time (JIT) compilation, the library mitigates the traditional performance overhead associated with per-example gradient clipping and noise addition. ### Core Components and Advanced Algorithms * The toolkit provides fundamental building blocks for implementing standard DP algorithms like DP-SGD and DP-FTRL, including specialized modules for data batch construction. * It supports state-of-the-art methods such as DP matrix factorization, which improves performance by injecting correlated noise across training iterations. * Features like micro-batching and padding are included to handle the massive, variable-sized batches often required to achieve an optimal balance between privacy and model utility. ### Verification and Privacy Auditing * JAX-Privacy incorporates rigorous privacy accounting based on Rényi Differential Privacy to provide precise tracking of privacy budgets. * The library includes tools for empirical auditing, allowing developers to validate their privacy guarantees through techniques like membership inference attacks and data poisoning. * The design ensures correctness in distributed settings, specifically focusing on consistent noise generation and gradient synchronization across clusters. JAX-Privacy 1.0 is a robust solution for researchers and engineers who need to deploy production-grade private models. Its modular architecture and integration with high-performance computing primitives make it a primary choice for training foundation models on sensitive datasets without compromising on scalability or security.

Machine Learning AI Differential Privacy Jax+5

googleNov 6, 2025Original article

Introducing Nested Learning: A new ML paradigm for continual learning (opens in new tab)

Google Research has introduced Nested Learning, a paradigm that treats machine learning models as systems of interconnected, multi-level optimization problems rather than separate architectures and training rules. By unifying structure and optimization through varying update frequencies, this approach aims to mitigate "catastrophic forgetting," the tendency for models to lose old knowledge when acquiring new skills. The researchers validated this framework through "Hope," a self-modifying architecture that outperforms current state-of-the-art models in long-context memory and language modeling. ### The Nested Learning Paradigm This framework shifts the view of machine learning from a single continuous process to a set of coherent, nested optimization problems. Each component within a model is characterized by its own "context flow"—the specific set of information it learns from—and its own update frequency. * The paradigm argues that architecture (structure) and optimization (training rules) are fundamentally the same concept, differing only by their level of computational depth and update rates. * Associative memory is used as a core illustrative concept, where the training process (backpropagation) is modeled as a system mapping data points to local error values. * By defining an update frequency rate for each component, researchers can order these problems into "levels," allowing for a more unified and efficient learning system inspired by the human brain's neuroplasticity. ### Deep Optimizers and Refined Objectives Nested Learning provides a principled way to improve standard optimization algorithms by viewing them through the lens of associative memory modules. * Existing momentum-based optimizers often rely on simple dot-product similarity, which fails to account for how different data samples relate to one another. * By replacing these simple similarities with standard loss metrics, such as L2 regression loss, the researchers derived new formulations for momentum that are more resilient to imperfect or noisy data. * This approach turns the optimizer itself into a deeper learning component with its own internal optimization objective. ### Continuum Memory Systems and the "Hope" Architecture The paradigm addresses the limitations of Large Language Models (LLMs), which are often restricted to either their immediate input window or static pre-trained knowledge. * The researchers developed "Hope," a proof-of-concept architecture that utilizes multi-time-scale updates for its internal components. * While standard Transformers act primarily as short-term memory, the Nested Learning approach allows for "continuum memory" that manages long-context information more effectively. * Experimental results show that this self-modifying architecture achieves superior performance in language modeling compared to existing state-of-the-art models. By recognizing that every part of a model is essentially an optimizer operating at a different frequency, Nested Learning offers a path toward AI that can adapt to new experiences in real-time. This structural shift moves away from the "static pre-training" bottleneck and toward systems capable of true human-like neuroplasticity and lifelong learning.

Machine Learning Deep Learning Transformer Optimization Algorithms+4

googleNov 5, 2025Original article

DS-STAR: A state-of-the-art versatile data science agent (opens in new tab)

DS-STAR is an advanced autonomous data science agent developed to handle the complexity and heterogeneity of real-world data tasks, ranging from statistical analysis to visualization. By integrating a specialized file analysis module with an iterative planning and verification loop, the system can interpret unstructured data and refine its reasoning steps dynamically based on execution feedback. This architecture allows DS-STAR to achieve state-of-the-art performance on major industry benchmarks, effectively bridging the gap between natural language queries and executable, verified code. ## Comprehensive Data File Analysis The framework addresses a major limitation of current agents—the over-reliance on structured CSV files—by implementing a dedicated analysis stage for diverse data formats. * The system automatically scans a directory to extract context from heterogeneous formats, including JSON, unstructured text, and markdown files. * A Python-based analysis script generates a textual summary of the data structure and content, which serves as the foundational context for the planning phase. * This module ensures the agent can navigate complex, multi-file environments where critical information is often spread across non-relational sources. ## Iterative Planning and Verification Architecture DS-STAR utilizes a sophisticated loop involving four specialized roles to mimic the workflow of a human expert conducting sequential analysis. * **Planner and Coder:** A Planner agent establishes high-level objectives, which a Coder agent سپس translates into executable Python scripts. * **LLM-based Verification:** A Verifier agent acts as a judge, assessing whether the generated code and its output are sufficient to solve the problem or if the reasoning is flawed. * **Dynamic Routing:** If the Verifier identifies gaps, a Router agent guides the refinement process by adding new steps or correcting errors, allowing the cycle to repeat for up to 10 rounds. * **Intermediate Review:** The agent reviews intermediate results before proceeding to the next step, similar to how data scientists use interactive environments like Google Colab. ## Benchmarking and State-of-the-Art Performance The effectiveness of the DS-STAR framework was validated through rigorous testing against existing agents like AutoGen and DA-Agent. * The agent secured the top rank on the public DABStep leaderboard, raising accuracy from 41.0% to 45.2% compared to previous best-performing models. * Performance gains were consistent across other benchmarks, including KramaBench (39.8% to 44.7%) and DA-Code (37.0% to 38.5%). * DS-STAR showed a significant advantage in "hard" tasks—those requiring the synthesis of information from multiple, varied data sources—demonstrating its superior versatility in complex environments. By automating the time-intensive tasks of data wrangling and verification, DS-STAR provides a robust template for the next generation of AI assistants. Organizations looking to scale their data science capabilities should consider adopting iterative agentic workflows that prioritize multi-format data understanding and self-correcting execution loops.

Machine Learning AI Large Language Models Python+4

netflixNov 4, 2025Original article

Netflix's Metaflow Spin: Faster ML Development | Netflix TechBlog (opens in new tab)

Netflix has introduced Spin, a new functionality within the Metaflow framework designed to significantly accelerate the iterative development cycle for ML and AI workflows. By bridging the gap between the interactive speed of notebooks and the production-grade reliability of versioned workflows, Spin allows developers to experiment with stateful increments without the latency of full restarts. This enhancement ensures that the "prototype to production" pipeline remains fluid while maintaining the deterministic execution and explicit state management that Metaflow provides at scale. ### The Nature of ML and AI Iteration * ML and AI development is distinct from traditional software engineering because it involves large, mutable datasets and computationally expensive, stochastic processes. * State management is a primary concern in this domain, as reloading data or recomputing transformations for every minor code change creates a prohibitively slow feedback loop. * While notebooks like Jupyter or Marimo excel at preserving in-memory state for fast exploration, they often lead to "hidden state" problems and non-deterministic results due to out-of-order cell execution. ### Metaflow as a State-Aware Framework * Metaflow uses the `@step` decorator to define checkpoint boundaries where the framework automatically persists all instance variables as versioned artifacts. * The framework’s `resume` command allows developers to restart execution from a specific step, cloning previous state to avoid recomputing successful upstream tasks. * This architecture addresses notebook limitations by ensuring execution order is explicit and deterministic while making the state fully discoverable and versioned. ### Introducing Spin for Rapid Development * Spin is a new feature introduced in Metaflow 2.19 that further reduces the friction of the iterative development loop. * It aims to provide the near-instant feedback of a notebook environment while operating within the structure of a production-ready Metaflow workflow. * The tool helps developers manage the stateful nature of ML development, allowing for quick, incremental experimentation without losing continuity between code iterations. To improve data science productivity and reduce "waiting time" during the development phase, engineering teams should look to adopt Metaflow 2.19 and integrate Spin into their experimentation workflows.

Machine Learning AI Data Science Workflow Orchestration+3

googleNov 4, 2025Original article

Forecasting the future of forests with AI: From counting losses to predicting risk (opens in new tab)

Research from Google DeepMind and Google Research introduces ForestCast, a deep learning-based framework designed to transition forest management from retrospective loss monitoring to proactive risk forecasting. By utilizing vision transformers and pure satellite data, the team has developed a scalable method to predict future deforestation that matches or exceeds the accuracy of traditional models dependent on inconsistent manual inputs. This approach provides a repeatable, future-proof benchmark for protecting biodiversity and mitigating climate change on a global scale. ### Limitations of Traditional Forecasting * Existing state-of-the-art models rely on specialized geospatial maps, such as infrastructure development, road networks, and regional economic indicators. * These traditional inputs are often "patchy" and inconsistent across different countries, requiring manual assembly that is difficult to replicate globally. * Manual data sources are not future-proof; they tend to go out of date quickly with no guarantee of regular updates, unlike continuous satellite streams. ### A Scalable Pure-Satellite Architecture * The ForestCast model adopts a "pure satellite" approach, using only raw inputs from Landsat and Sentinel-2 satellites. * The architecture is built on vision transformers (ViTs) that process an entire tile of pixels in a single pass to capture critical spatial context and landscape-level trends. * The model incorporates a satellite-derived "change history" layer, which identifies previously deforested pixels and the specific year the loss occurred. * By avoiding socio-political or infrastructure maps, the method can be applied consistently to any region on Earth, allowing for meaningful cross-regional comparisons. ### Key Findings and Benchmark Release * Research indicates that "change history" is the most information-dense input; a model trained on this data alone performs almost as well as those using raw multi-spectral data. * The model successfully predicts tile-to-tile variation in deforestation amounts and identifies the specific pixels most likely to be cleared next. * Google has released the training and evaluation data as a public benchmark dataset, focusing initially on Southeast Asia to allow the machine learning community to verify and improve upon the results. The release of ForestCast provides a template for scaling predictive modeling to Latin America, Africa, and boreal latitudes. Conservationists and policymakers should utilize these forecasting tools to move beyond counting historical losses and instead direct resources toward "frontline" areas where the model identifies imminent risk of habitat conversion.

Machine Learning AI Deep Learning Predictive Modeling+4

googleNov 3, 2025Original article

Exploring a space-based, scalable AI infrastructure system design (opens in new tab)

Project Suncatcher is a Google moonshot initiative aimed at scaling machine learning infrastructure by deploying solar-powered satellite constellations equipped with Tensor Processing Units (TPUs). By leveraging the nearly continuous energy of the sun in specific orbits and utilizing high-bandwidth free-space optical links, the project seeks to bypass the resource constraints of terrestrial data centers. Early research suggests that a modular, tightly clustered satellite design can achieve the necessary compute density and communication speeds required for modern AI workloads. ### Data-Center Bandwidth via Optical Links * To match terrestrial performance, inter-satellite links must support tens of terabits per second using multi-channel dense wavelength-division multiplexing (DWDM) and spatial multiplexing. * The system addresses signal power loss (the link budget) by maintaining satellites in extremely close proximity—kilometers or less—compared to traditional long-range satellite deployments. * Initial bench-scale demonstrations have successfully achieved 800 Gbps each-way transmission (1.6 Tbps total) using a single transceiver pair, validating the feasibility of high-speed optical networking. ### Orbital Mechanics of Compact Constellations * The proposed system utilizes a sun-synchronous low-earth orbit (LEO) at an altitude of approximately 650 km to maximize solar exposure and minimize the weight of onboard batteries. * Researchers use Hill-Clohessy-Wiltshire equations and JAX-based differentiable models to manage the complex gravitational perturbations and atmospheric drag affecting satellites flying in tight 100–200m formations. * Simulations of 81-satellite clusters indicate that only modest station-keeping maneuvers are required to maintain stable, "free-fall" trajectories within the orbital plane. ### Hardware Resilience in Space Environments * The project specifically tests Google’s Trillium (v6e) Cloud TPUs to determine if terrestrial AI accelerators can survive the radiation found in LEO. * Hardware is subjected to 67MeV proton beams to analyze the impact of Total Ionizing Dose (TID) and Single Event Effects (SEEs) on processing reliability. * Preliminary testing indicates promising results for the radiation tolerance of high-performance accelerators, suggesting that standard TPU architectures may be viable for orbital deployment with minimal modification. While still in the research and development phase, Project Suncatcher suggests that the future of massive AI scaling may involve shifting infrastructure away from terrestrial limits and toward modular, energy-rich orbital environments. Organizations should monitor the progress of free-space optical communication and radiation-hardened accelerators as these technologies will be the primary gatekeepers for space-based computation.

Machine Learning AI Ai Infrastructure Jax+5

netflixOct 25, 2025Original article

Post-Training Generative Recommenders with Advantage-Weighted Supervised Finetuning | by Netflix Technology Blog | Netflix TechBlog (opens in new tab)

Netflix is evolving its recommendation systems by moving beyond simple behavior imitation toward generative recommenders that better align with true user preferences. While generative models like HSTU and OneRec effectively capture sequential user patterns, they often struggle to distinguish between habitual clicks and genuine satisfaction. To bridge this gap, Netflix developed Advantage-Weighted Supervised Fine-tuning (A-SFT), a post-training method that leverages noisy reward signals to refine model performance without the need for complex counterfactual data. ### The Shift to Generative Recommenders * Modern generative recommenders (GRs), such as HSTU and OneRec, utilize transformer architectures to treat recommendation as a sequential transduction task. * The models are typically trained using next-item prediction, where the system learns to imitate the chronological sequence of a user’s activities. * A significant drawback of this "behavior cloning" approach is that it captures external trends and noise rather than long-term user satisfaction, potentially recommending content the user finished but did not actually enjoy. ### Barriers to Reinforcement Learning in RecSys * Traditional post-training methods used in Large Language Models, such as Proximal Policy Optimization (PPO) or Direct Preference Optimization (DPO), require counterfactual feedback that is difficult to obtain in recommendation contexts. * Because user sequences span weeks or years, it is impractical to generate and test hypothetical, counterfactual experiences for real-time user validation. * Reward signals in recommendation systems are inherently noisy; for instance, high watch time might indicate interest, but it can also be a result of external circumstances, making it an unreliable metric for optimization. ### Advantage-Weighted Supervised Fine-tuning (A-SFT) * A-SFT is a hybrid approach that sits between offline reinforcement learning and standard supervised fine-tuning. * The algorithm incorporates an advantage function to weight training examples, allowing the model to prioritize actions that lead to higher rewards while filtering out noise from the reward model. * This method is specifically designed to handle high-variance reward signals, using them as directional guides rather than absolute truth, which prevents the model from over-exploiting inaccurate data. * Benchmarks against other representative methods show that A-SFT achieves superior alignment between the generative recommendation policy and the underlying reward model. For organizations managing large-scale recommendation engines, A-SFT offers a practical path to implementing post-training improvements. By focusing on advantage-weighted signals, developers can improve recommendation quality using existing implicit feedback—like watch time and clicks—without the infrastructure hurdles of online reinforcement learning.

Machine Learning AI Transformer Recommendation Systems+4

dropboxOct 22, 20253 min readCurated summary

Half-Quadratic Quantization of large machine learning models

Half-Quadratic Quantization (HQQ) is a calibration-free method for compressing large machine learning models while retaining quality comparable to calibration-based techniques such as GPTQ and AWQ. It minimizes weight reconstruction error rather than activation error and uses a sparsity-promoting \(l_p\) loss to better handle outliers. Because HQQ relies on closed-form alternating updates instead of gradient-based optimization, it can quantize models dramatically faster—reportedly processing Llama-2-70B in under five minutes. ## Why Quantization Matters - Large language models require substantial memory for training and inference. - Methods such as bitsandbytes, GPTQ, and AWQ make models like Llama-2 usable on consumer GPUs. - Weight-only quantization approaches fall into two groups: - **Calibration-free methods**, such as bitsandbytes, use only model weights. - **Calibration-based methods**, such as GPTQ and AWQ, use external datasets. - Calibration-based approaches can provide better quality but: - Their results may depend on calibration-data bias. - Calibration can be computationally expensive for very large models. ## HQQ’s Quantization Objective - Standard quantization can significantly distort weights, particularly outliers with unusually large values. - GPTQ and AWQ reduce the effect of these distortions by minimizing layer-output or activation error using calibration data. - HQQ instead minimizes reconstruction error directly in the weights. - It uses a sparsity-promoting \(l_p\) loss, especially with \(p<1\), to model heavy-tailed outlier errors more effectively than squared error. - Quantization is defined using: - A scale \(s\) - A zero-point \(z\) - A quantization operator \(Q_{z,s}(W)=\text{round}(W/s+z)\) - A dequantization operator \(Q^{-1}_{z,s}(W_q)=s(W_q-z)\) - HQQ fixes the scale and optimizes the zero-point, simplifying the optimization problem. ## Half-Quadratic Optimization - Since the \(l_p\) objective with \(p<1\) is non-convex, HQQ introduces an auxiliary error variable \(W_e\). - The resulting problem is solved through alternating optimization: - Update \(W_e\) while holding \(z\) fixed. - Update \(z\) while holding \(W_e\) fixed. - Increase a positive penalty parameter \(\beta\) by a factor \(\kappa\) each iteration. - This decomposition turns the original difficult problem into simpler sub-problems with closed-form solutions. ## Solving the Sub-Problems - The \(W_e\) update is a proximal operation. - For \(l_1\) regularization, it corresponds to soft thresholding. - HQQ uses a generalized soft-thresholding operator for \(0\leq p\leq1\): \[ \text{shrink}_{l_p}(x,\beta) =\text{sign}(x)\,\text{relu}\left(|x|-\frac{|x|^{p-1}}{\beta}\right) \] - The zero-point update: - Recomputes quantized weights using the current zero-point. - Calculates the difference between quantized weights and corrected original weights. - Sets the new zero-point to the average over the quantization grouping axis. - The implementation optimizes the inverse scale \(1/s\), which is more numerically stable in half-precision arithmetic. ## Speed and Practical Advantages - HQQ uses closed-form updates rather than gradients or automatic differentiation. - Quantization can run in inference mode with half-precision arithmetic. - The solver typically converges in only a few iterations. - In contrast, AdamW with PyTorch autograd may require thousands of iterations and fails when using \(p<1\). - The article reports HQQ as: - More than 100 times faster than autograd for quantizing Llama-2-7B. - More than 50 times faster than GPTQ for Llama-2-70B. - Capable of quantizing the largest models in only a few minutes. - A 2-bit HQQ version of Llama-2-70B reportedly outperforms full-precision Llama-2-13B at a comparable memory footprint. HQQ is therefore presented as a practical alternative to calibration-based quantization: it combines calibration-free operation and very high speed with competitive compression quality, making rapid experimentation and deployment of large models more feasible.

Read original

Machine Learning Large Language Models Inference Optimization Model Quantization+4

googleOct 16, 2025Original article

Solving virtual machine puzzles: How AI is optimizing cloud computing (opens in new tab)

Google researchers have developed LAVA, a scheduling framework designed to optimize virtual machine (VM) allocation in large-scale data centers by accurately predicting and adapting to VM lifespans. By moving beyond static, one-time predictions toward a "continuous re-prediction" model based on survival analysis, the system significantly improves resource efficiency and reduces fragmentation. This approach allows cloud providers to solve the complex "bin packing" problem more effectively, leading to better capacity utilization and easier system maintenance. ### The Challenge of Long-Tailed VM Distributions * Cloud workloads exhibit a extreme long-tailed distribution: while 88% of VMs live for less than an hour, these short-lived jobs consume only 2% of total resources. * The rare VMs that run for 30 days or longer account for a massive fraction of compute resources, meaning their placement has a disproportionate impact on host availability. * Poor allocation leads to "resource stranding," where a server's remaining capacity is too small or unbalanced to host new VMs, effectively wasting expensive hardware. * Traditional machine learning models that provide only a single prediction at VM creation are often fragile, as a single misprediction can block a physical host from being cleared for maintenance or new tasks. ### Continuous Re-prediction via Survival Analysis * Instead of predicting a single average lifetime, LAVA uses an ML model to generate a probability distribution of a VM's expected duration. * The system employs "continuous re-prediction," asking how much longer a VM is expected to run given how long it has already survived (e.g., a VM that has run for five days is assigned a different remaining lifespan than a brand-new one). * This adaptive approach allows the scheduling logic to automatically correct for initial mispredictions as more data about the VM's actual behavior becomes available over time. ### Novel Scheduling and Rescheduling Algorithms * **Non-Invasive Lifetime Aware Scheduling (NILAS):** Currently deployed on Google’s Borg cluster manager, this algorithm ranks potential hosts by grouping VMs with similar expected exit times to increase the frequency of "empty hosts" available for maintenance. * **Lifetime-Aware VM Allocation (LAVA):** This algorithm fills resource gaps on hosts containing long-lived VMs with jobs that are at least an order of magnitude shorter. This ensures the short-lived VMs exit quickly without extending the host's overall occupation time. * **Lifetime-Aware Rescheduling (LARS):** To minimize disruptions during defragmentation, LARS identifies and migrates the longest-lived VMs first while allowing short-lived VMs to finish their tasks naturally on the original host. By integrating survival-analysis-based predictions into the core logic of data center management, cloud providers can transition from reactive scheduling to a proactive model. This system not only maximizes resource density but also ensures that the physical infrastructure remains flexible enough to handle large, resource-intensive provisioning requests and essential system updates.

Machine Learning AI Cloud Computing Virtual Machines+5

googleOct 15, 2025Original article

Using AI to identify genetic variants in tumors with DeepSomatic (opens in new tab)

DeepSomatic is an AI-powered tool developed by Google Research to identify cancer-related mutations by analyzing a tumor's genetic sequence with higher accuracy than current methods. By leveraging convolutional neural networks (CNNs), the model distinguishes between inherited genetic traits and acquired somatic variants that drive cancer progression. This flexible tool supports multiple sequencing platforms and sample types, offering a critical resource for clinicians and researchers aiming to personalize cancer treatment through precision medicine. ## Challenges in Somatic Variant Detection * Somatic variants are genetic mutations acquired after birth through environmental exposure or DNA replication errors, making them distinct from the germline variants found in every cell of a person's body. * Detecting these mutations is technically difficult because tumor samples are often heterogeneous, containing a diverse set of variants at varying frequencies. * Sequencing technologies often introduce small errors that can be difficult to distinguish from actual somatic mutations, especially when the mutation is only present in a small fraction of the sampled cells. ## CNN-Based Variant Calling Architecture * DeepSomatic employs a method pioneered by DeepVariant, which involves transforming raw genetic sequencing data into a set of multi-channel images. * These images represent various data points, including alignment along the chromosome, the quality of the sequence output, and other technical variables. * The convolutional neural network processes these images to differentiate between three categories: the human reference genome, non-cancerous germline variants, and the somatic mutations driving tumor growth. * By analyzing tumor and non-cancerous cells side-by-side, the model effectively filters out sequencing artifacts that might otherwise be misidentified as mutations. ## System Versatility and Application * The model is designed to function in multiple modes, including "tumor-normal" (comparing a biopsy to a healthy sample) and "tumor-only" mode, which is vital for blood cancers like leukemia where isolating healthy cells is difficult. * DeepSomatic is platform-agnostic, meaning it can process data from all major sequencing technologies and adapt to different types of sample processing. * The tool has demonstrated the ability to generalize its learning to various cancer types, even those not specifically included in its initial training sets. ## Open-Source Contributions to Precision Medicine * Google has made the DeepSomatic tool and the CASTLE dataset—a high-quality training and evaluation set—openly available to the global research community. * This initiative is part of a broader effort to use AI for early detection and advanced research in various cancers, including breast, lung, and gynecological cancers. * The release aims to accelerate the development of personalized treatment plans by providing a more reliable way to identify the specific genetic drivers of an individual's disease. By providing a more accurate and adaptable method for variant calling, DeepSomatic helps researchers pinpoint the specific drivers of a patient's cancer. This tool represents a significant advancement in deep learning for genomics, potentially shortening the path from biopsy to targeted therapeutic intervention.

Machine Learning AI Deep Learning Genomics+5

googleOct 14, 2025Original article

Coral NPU: A full-stack platform for Edge AI (opens in new tab)

Coral NPU is a new full-stack, open-source platform designed to bring advanced AI directly to power-constrained edge devices and wearables. By prioritizing a matrix-first hardware architecture and a unified software stack, Google aims to overcome traditional bottlenecks in performance, ecosystem fragmentation, and data privacy. The platform enables always-on, low-power ambient sensing while providing developers with a flexible, RISC-V-based environment for deploying modern machine learning models. ## Overcoming Edge AI Constraints * The platform addresses the "performance gap" where complex ML models typically exceed the power, thermal, and memory budgets of battery-operated devices. * It eliminates the "fragmentation tax" by providing a unified architecture, moving away from proprietary processors that require costly, device-specific optimizations. * On-device processing ensures a high standard of privacy and security by keeping personal context and data off the cloud. ## AI-First Hardware Architecture * Unlike traditional chips, this architecture prioritizes the ML matrix engine over scalar compute to optimize for efficient on-device inference. * The design is built on RISC-V ISA compliant architectural IP blocks, offering an open and extensible reference for system-on-chip (SoC) designers. * The base design delivers performance in the 512 giga operations per second (GOPS) range while consuming only a few milliwatts of power. * The architecture is tailored for "always-on" use cases, making it ideal for hearables, AR glasses, and smartwatches. ## Core Architectural Components * **Scalar Core:** A lightweight, C-programmable RISC-V frontend that manages data flow using an ultra-low-power "run-to-completion" model. * **Vector Execution Unit:** A SIMD co-processor compliant with the RISC-V Vector instruction set (RVV) v1.0 for simultaneous operations on large datasets. * **Matrix Execution Unit:** A specialized engine using quantized outer product multiply-accumulate (MAC) operations to accelerate fundamental neural network tasks. ## Unified Developer Ecosystem * The platform is a C-programmable target that integrates with modern compilers such as IREE and TFLM (TensorFlow Lite Micro). * It supports a wide range of popular ML frameworks, including TensorFlow, JAX, and PyTorch. * The software toolchain utilizes MLIR and the StableHLO dialect to facilitate the transition from high-level models to hardware-executable code. * Developers have access to a complete suite of tools, including a simulator, custom kernels, and a general-purpose MLIR compiler. SoC designers and ML developers looking to build the next generation of wearables should leverage the Coral NPU reference architecture to balance high-performance AI with extreme power efficiency. By utilizing the open-source documentation and RISC-V-based tools, teams can significantly reduce the complexity of deploying private, always-on ambient sensing.

Machine Learning AI Coral Npu Edge Ai+5

googleSep 30, 2025Original article

Introducing interactive on-device segmentation in Snapseed (opens in new tab)

Google has introduced a new "Object Brush" feature in Snapseed that enables intuitive, real-time selective photo editing through a novel on-device segmentation technology. By leveraging a high-performance interactive AI model, users can isolate complex subjects with simple touch gestures in under 20 milliseconds, bridging the gap between professional-grade editing and mobile convenience. This breakthrough is achieved through a sophisticated teacher-student training architecture that prioritizes both pixel-perfect accuracy and low-latency performance on consumer hardware. ### High-Performance On-Device Inference * The system is powered by the Interactive Segmenter model, which is integrated directly into the Snapseed "Adjust" tool to facilitate immediate object-based modifications. * To ensure a fluid user experience, the model utilizes the MediaPipe framework and LiteRT’s GPU acceleration to process selections in less than 20ms. * The interface supports dynamic refinement, allowing users to provide real-time feedback by tracing lines or tapping to add or subtract specific areas of an image. ### Teacher-Student Model Distillation * The development team first created "Interactive Segmenter: Teacher," a large-scale model fine-tuned on 30,000 high-quality, pixel-perfect manual annotations across more than 350 object categories. * Because the Teacher model’s size and computational requirements are prohibitive for mobile use, researchers developed "Interactive Segmenter: Edge" through knowledge distillation. * This distillation process utilized a dataset of over 2 million weakly annotated images, allowing the smaller Edge model to inherit the generalization capabilities of the Teacher model while maintaining a footprint suitable for mobile devices. ### Training via Synthetic User Prompts * To make the model universally capable across all object types, the training process uses a class-agnostic approach based on the Big Transfer (BiT) strategy. * The model learns to interpret user intent through "prompt generation," which simulates real-world interactions such as random scribbles, taps, and lasso (box) selections. * During training, both the Teacher and Edge models receive identical prompts—such as red foreground scribbles and blue background scribbles—to ensure the student model learns to produce high-quality masks even from imprecise user input. This advancement significantly lowers the barrier to entry for complex photo manipulation by moving heavy-duty AI processing directly onto the mobile device. Users can expect a more responsive and precise editing experience that handles everything from fine-tuning a subject's lighting to isolating specific environmental elements like clouds or clothing.

Machine Learning AI Computer Vision On Device Ai+5