Loading…
Topic
622 posts about Machine Learning. Every summary links to the original.
See the idea
A walkthrough of how a 3D CT model turns a voxel volume into a structured diagnostic report: isotropic resampling, 3D patch tokens, spatial encoding, and language-model reasoning.
The Good Stuff
Paid picks we use ourselves. Links carry a referral code — a subscription through one earns us a commission at no extra cost to you.
Newest first
Golden Gate Claude presents a research demonstration of interpretability work on Claude 3 Sonnet. Researchers identified millions of activated concepts, or “features,” including a specific combination of neurons associated with the Golden Gate Bridge, and adjusted that feature’s activation strength. Increasing it caused responses to mention the bridge across unrelated prompts, such as spending $10 or writing a love story; the demonstration was available online for 24 hours and is no longer available. The post distinguishes this intervention from verbal role-play, system prompts, and traditional fine-tuning, describing it as a precise change to internal activations. It also says similar techniques could modify safety-related features connected to dangerous code, criminal activity, or deception, potentially supporting safer models.
Real-world financial natural language processing datasets often overrepresent common events like earnings while underrepresenting rarer occurrences such as credit-rating changes. To address this imbalance, an iterative generation pipeline produced over 500,000 unique headlines across 13 categories using NVIDIA NeMo Data Designer, NeMo Curator, and Nemotron 3 Nano. Instead of relying on a single large batch that yielded 65 percent near-duplicates, the workflow executes iterative cycles of category-weighted generation, rule-based filtering, global semantic deduplication, and dynamic category reweighting. Farthest-from-centroid few-shot example selection guided subsequent rounds toward novel semantic spaces across 82 total iterations. Executed over approximately six days on an eight-way NVIDIA B200 node, the resulting corpus enabled fine-tuning compact student language models to achieve 95 percent of teacher model performance.
amazon.science
Reinforcement learning for agentic tasks requires updating model parameters against the precise token sequences generated during rollouts. Standard agent harnesses frequently modify conversation transcripts through message compaction and tool formatting, causing retokenization drift that degrades policy-gradient training. Turnstile addresses this issue as a Rust-based proxy situated between agent harnesses and inference backends such as SGLang. By intercepting OpenAI-compatible Chat Completions traffic, the proxy records exact token IDs, log probabilities, loss masks, weight versions, and mixture-of-experts routing traces at the moment of generation. Validation across text-based coding and multimodal computer-use agents demonstrated steady reinforcement learning improvements without requiring modifications to agent harnesses.
Anthropic argues that AI could have an impact comparable to the industrial and scientific revolutions, potentially within the coming decade, because scaling laws link increased training computation with predictable capability improvements. It says progress may continue toward systems matching or exceeding human performance across many intellectual tasks, while society faces disruption and competitive pressure to deploy systems that are not reliably helpful, honest, and harmless. The company’s safety strategy is empirically driven and multifaceted, emphasizing scalable supervision, mechanistic interpretability, process-oriented learning, and research into how systems learn and generalize. It also studies capability evaluation, red teaming, harmful behavior, economic effects, and policy implications. The stated goal is a broad portfolio of safety work that can adapt as evidence clarifies which future scenario is unfolding, including less optimistic possibilities.
Uber describes how it uses early chargeback signals, especially TC40 reports, to identify payment-fraud patterns before conventional chargebacks arrive. These notifications come from payment service providers through fraud reports and are ingested from external vendors via SFTP, third-party APIs, or webhooks, then validated, normalized, published to Apache Kafka, and stored in Apache Hive. A streaming pipeline maps signals to internal transactions and users, derives near-real-time features, and sends them to machine-learning models and a rule engine that determines risk actions. Redis TTL caching filters duplicate references, while Kafka propagates signals to downstream systems and Hive retains them for batch processing. Since launch, the project has processed more than 2 million additional signals, and TC40 signals arrive on average 4–5 days earlier than chargeback signals, although the source notes that TC40 data alone does not prevent chargebacks.
Agentic AI workflows and reinforcement learning rely heavily on sequential CPU tasks, such as tool calls, sandboxed code execution, data processing, and environment rollouts between GPU steps. When CPU performance degrades under full socket load, training time inflates, serving latency grows, and GPU KV-caches face eviction. The NVIDIA Vera CPU addresses these bottlenecks by featuring 88 Olympus cores on a monolithic compute die, paired with power-efficient LPDDR5x memory and the Scalable Coherency Fabric. These cores incorporate a neural branch predictor, a 10-wide decode front end, deep out-of-order execution, and NVIDIA Spatial Multithreading. Consequently, Vera achieves 1.8x faster core execution, 40% lower peak loaded latency compared to x86 CPUs, and up to 1.2 TB/s total memory bandwidth, preventing costly KV-cache recomputations.
Humanoid robotics development suffers from fragmented tooling, siloed software ecosystems, and incompatible data formats that hinder reproducible workflows. NVIDIA introduced the Isaac GR00T Development Platform, an open-source, modular humanoid development environment that connects simulation setup, demonstration capture, policy training, evaluation, and hardware deployment. Central to this platform is the Isaac GR00T 1.7 vision-language-action model, which features a 3-billion-parameter base checkpoint licensed under Apache 2.0. The architecture adopts a Cosmos-Reason2-2B backbone and supports full pipeline export to ONNX and TensorRT for deployment on Jetson Thor. In a reference pick-and-place manipulation task, developers capture teleoperated demonstrations with Isaac Teleop, convert trajectories into LeRobot format, post-train GR00T 1.7, and evaluate the closed-loop policy inside Isaac Lab-Arena.
Massive MIMO deployments often fail to achieve theoretical spectral efficiency because CPU compute constraints force algorithmic compromises in user tracking, interference mitigation, and pairing. NVIDIA AI Aerial addresses this gap using a GPU-accelerated, AI-native radio access network architecture. This approach enables dense mathematical models across Layer 1 and Layer 2 workloads, including machine-learning beamforming and deep reinforcement learning link adaptation. In 64T64R simulations, AI beamforming achieved up to 1.62x higher throughput at 32 layers compared to regularized Zero Forcing baselines. The platform also unifies physical and MAC layer coordination while allowing dynamic allocation of spare GPU compute to edge AI inference.
Industrial machinery produces high volumes of alarms and sensor readings that exceed manual triage capacity. To address this bottleneck, an end-to-end GPU-accelerated AI agent is designed using NVIDIA NeMo libraries, Nemotron open models, and the NVIDIA OpenShell runtime behind a single HTTP endpoint. When triggered with an alarm payload and sensor frame, the agent gathers historical context using cuDF, cuVS, and NeMo Retriever, queries data warehouses via Text-2-SQL, and executes specialist signal checks with tools like cuFFT and cuML. Nemotron 3 Nano manages orchestration tasks while Nemotron 3 Super synthesizes the gathered evidence. The workflow outputs a structured package containing an observation, root-cause hypothesis, remedy, and recommended action within a seconds-long latency budget.
Large-scale LLM training jobs spanning thousands of GPUs frequently encounter transient hardware interruptions that bottleneck tightly coupled tensor parallel groups. Nonuniform Tensor Parallelism addresses this bottleneck by dynamically adapting the tensor parallelism degree within affected scale-up domains rather than dropping entire data replicas. When a GPU drops out, the system reconfigures the remaining functional devices to process larger shards and resumes execution from the latest checkpoint. To prevent the downsized replica from stalling global synchronization, advanced rack designs temporarily power-boost the active GPUs to increase clock frequencies and match baseline throughput. Overlapping the necessary tensor resharding with backward computation and parameter synchronization reduces overhead on healthy replicas to under one percent.
Classic intrusion detection signatures suffer from an exposure window between zero-day discovery and rule distribution, leaving slight exploit variations unflagged. Cisco Talos addressed this by embedding SnortML into Snort 3 to perform local, sub-millisecond machine learning inference alongside signature checks. SnortML employs an LSTM preceded by a byte-embedding layer and XNNPACK acceleration to score URI queries and POST bodies for vulnerabilities like SQL injection, XSS, and command injection. Running both detection mechanisms concurrently provides independent error profiles and boosts detection confidence when both fire. The broader defense landscape is shifting toward connecting these packet-level ML detectors with agentic AI pipelines and supervised feedback loops.
ClickHouse 26.6 is a release marking the 10-year anniversary of ClickHouse becoming open source, with 56 new features, 79 performance optimizations, and 366 bug fixes. Hypothetical skip indexes let users evaluate set indexes in a session without building them; on a 487.24-million-row UK property dataset, the example estimated skip ratios of 77.0% at granularity 128 and 92.2% at granularity 1. The release overhauls dependencies between refreshable materialized views, addressing schedule drift that could make dependent stages lag or miss a refresh cycle. It also introduces experimental continuous queries, the aiEmbed function, embedding quantization to 8 bits and below, and workload-level memory reservations.
NVIDIA Confidential Computing secures enterprise data, proprietary model weights, and the model itself during active AI inference. The architecture integrates silicon-level private signing keys fused during manufacturing, encrypted multi-GPU NVLink, and remote attestation through the NVIDIA Remote Attestation Service alongside CPU Trusted Execution Environments. To mitigate overhead from secure work submission latency and encrypted host-to-device transfers, software optimizations incorporate CC-safe autotuner timing in FlashInfer, asynchronous copy workers in SGLang, and piecewise CUDA graph replay. Benchmarks on an HGX B300 system running the Qwen 3.5 397B-A17B model at FP8 precision show inference performance reaches up to 98% of non-confidential baselines across tested concurrency levels.
Databricks AI describes GPU reliability as a fleet-scale problem involving crashed jobs, silent slowdowns, and numerical corruption during distributed training. Its approach combines stress testing with demanding reinforcement learning, agentic coding, and document intelligence workloads, plus gpu-monitor, a multi-stage service that checks nodes during bootstrap, continuously while active, and periodically between workloads. The checks cover compute burn-in, peer connectivity, NCCL correctness and bandwidth, RDMA, ECC and HBM memory, PCIe integrity, throttling, thermal conditions, XID errors, and fabric health. A production incident showed why cumulative InfiniBand port downtime matters: a single outage exceeded the lower-level NCCL_IB_TIMEOUT, killing a collective before the longer PyTorch watchdog timeout. Nodes that fail checks are quarantined, while inter-node probes sweep message sizes from 8 bytes to 2 GiB and apply latency or BusBW criteria; the system is continuously expanded as new failure modes emerge.
Claude Fable 5 and Claude Mythos 5 are presented as a joint launch of the same underlying model, with Fable 5 intended for general use and Mythos 5 initially reserved for selected cyberdefenders and infrastructure providers. Fable 5 uses conservative safeguards that route some sensitive queries to Claude Opus 4.8, while Mythos 5 lifts safeguards in selected areas through Project Glasswing and a planned trusted-access program. The post reports strong results in software engineering, knowledge work, vision, long-context tasks, cybersecurity, protein design, molecular biology, and genomics, including a reported 50-million-line Ruby migration completed in a day. Both models are priced at $10 per million input tokens and $50 per million output tokens, with Fable 5 available through the Claude API and staged subscription access.
Language model agents often fail across long-horizon workflows, repeat tool-call errors, and output invalid schemas when relying solely on prompting or retrieval. Reinforcement learning with verifiable rewards provides a practical training signal for domain-specific tasks by scoring model rollouts against deterministic verifiers such as schema checkers, unit tests, and execution sandboxes. Teams can implement group relative policy optimization to sample multiple trajectories per prompt and update policy weights based on relative performance within the group. Establishing environments with explicit harness, state, and verifier definitions allows agents to learn multi-step tool use, whereas static datasets fail to capture sequential dependencies. Inspecting checkpoint rollouts and testing against held-out tasks prevents reward hacking while continuously turning operational failures into verifiable benchmark environments.
Retail and CPG forecasting now spans hundreds of thousands, sometimes far more, time series across fragmented channels, promotions, and short-lived products, making legacy methods and manual exception management difficult. Multi-model forecasting addresses this complexity by evaluating a range of techniques against actual data and selecting the best-performing model for each series, but enterprise-scale experiments require scarce forecasting and distributed-systems expertise. Released in 2024, Databricks’ open-source Many Model Forecasting (MMF) integrates more than 35 statistical, deep-learning, and foundation time-series models and runs on distributed Databricks compute. MMF Agent, built on Genie Code, guides users through data quality, series classification, compute configuration, forecasting, post-processing, and model selection, while Unity Catalog helps it use organizational data context. The workflow is intended to reduce setup from days to hours, improve targeting and accuracy, and make rigorous forecasting more accessible while remaining customizable for technical teams.
NVIDIA Omniverse NuRec reconstructs dynamic 3D environments from camera and lidar sensor data for autonomous vehicle simulation workflows, but high computational demands historically caused long reconstruction times. Engineers used NVIDIA Nsight Systems and Nsight Compute to systematically profile the PyTorch-based training pipeline and optimize its underlying CUDA kernels. Nsight Systems revealed GPU underutilization caused by numerous small kernels and blocking synchronization calls, leading to a fused interpolate kernel that accelerated execution from 4.184 milliseconds to 83.81 microseconds. Profiling with Nsight Compute uncovered resource mismatch in the renderBackward kernel, prompting the team to split it into dedicated camera and lidar implementations with tuned register and shared memory allocations. These adjustments raised kernel occupancy from approximately 15% to between 30% and 50% while decreasing the longest lidar kernel runtime from 31 milliseconds to 18 milliseconds.
Ramp describes an architecture for scaling risk operations as payment volume, payment rails, partners, and country coverage expand, rather than letting headcount grow with the business. Agents handle universal intake, gather business context, classify requests, and route work, while machine-learning models trained on millions of historical data points and approved policies make autonomous risk decisions. Operators can modify agent skills and tool configurations without engineering involvement, deploy changes in shadow mode, and rely on asynchronous execution, provider failover, and centralized observability. In payment-risk workflows, structured operator feedback produced a benchmark of more than 1,000 operations covering tool-call trajectories, operator alignment, and downstream outcomes, enabling separate evaluation of agents and policies. Ramp rolls agents out beside operators, then scales operation types and dollar volume using performance thresholds and capped exposure budgets.
Traditional large language models rely on next-token prediction trained over broad text sequences, excelling at conversational synthesis and code generation while struggling with forecasting future user behavior. Predicting intent inherently requires modeling decision-making under conditions of uncertainty, a capability not directly produced by general text pretraining. Dedicated behavioral foundation models address this limitation by training representations directly on proprietary, sensitive behavioral data, such as anonymous browsing sessions and product identifiers. By learning base representations of behavior, these models generalize to unseen predictive tasks, such as optimizing ad campaigns for products absent from the original training dataset. Implementing these behavioral models also necessitates privacy-centric machine learning techniques, ranging from differential privacy concepts to homomorphic machine learning.
The best engineering reads, once a week.
