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
NVIDIA announced the launch of NVIDIA Halos for Robotics, a full-stack functional safety system designed for industrial robots, humanoids, and autonomous mobile robots. The platform extends NVIDIA's functional safety architecture from autonomous vehicles into robotics, utilizing NVIDIA IGX Thor and the NVIDIA Holoscan Sensor Bridge as its hardware foundation. On top of this hardware, Halos OS provides the base Halos Core safety operating system and application blueprints, including an Outside-In Safety Blueprint that leverages external worksite cameras and safety decision modules. Platform hardware safety features include an isolated IEC 61508 SIL 3 capable Safety Island alongside extensive fault-detection mechanisms. To accelerate compliance, the accredited NVIDIA Halos AI Systems Inspection Lab assesses partner integrations before final third-party certification.
METRO Markets outgrew its self-hosted Impala and Kudu-based Hadoop stack as the marketplace expanded, with parallel stores, inconsistent reporting, and too much maintenance burdening a lean data team. They evaluated BigQuery, Snowflake, and ClickHouse against production queries, finding faster ClickHouse performance, including one response falling from four seconds to under 200 milliseconds, while its pricing supported more predictable costs. Airbyte moves CDC data from MySQL and Postgres into ClickHouse; ReplacingMergeTree handles deduplication during background merges, and BigQuery-hosted Google Analytics data enters through cloud storage and external tables. The resulting warehouse consolidates company data and supports seller analytics, credit risk features through Query API endpoints, observability, and emerging AI use cases, while teams query a shared dataset through compatible tools; planned Kafka and Flink integration would extend it to streaming workloads.
Data scientists connect statistics, programming, and domain knowledge to turn raw data into models, recommendations, and decisions with business consequences. The role now extends beyond classical modeling to large language models, generative AI applications, agentic systems, and production workflows involving deployment, monitoring, and retraining. Modern practice combines Python, SQL, and libraries such as scikit-learn, PyTorch, Spark, and MLflow with data engineering basics, communication, and domain expertise. Data scientists contribute across problem framing, data access, preparation, feature engineering, experimentation, deployment, and lifecycle monitoring, with governed access and lineage helping align training data with production data. The article concludes that AI assistants and agents can automate routine work, but human judgment remains necessary for trustworthy evaluation, business framing, and actionable recommendations.
MosaicLeaks examines privacy leakage in deep-research agents that combine private enterprise documents with external web retrieval. Queries that appear harmless individually can reveal sensitive facts when an observer combines them over multiple hops, producing intent, answer, or full-information leakage. The benchmark contains 1,001 multi-hop chains interleaving local and web questions, and compares baseline behavior with task-only reinforcement learning and Privacy-Aware Deep Research (PA-DR). Task-only training raises strict chain success from 48.7% to 59.3% but increases answer/full-information leakage from 34.0% to 51.7%, whereas PA-DR reaches 58.7% success with 9.9% leakage by rewarding stage-specific decisions and penalizing query-level privacy risk. The evaluation is controlled, using synthetic enterprise documents, a fixed web corpus, three company contexts, and one multi-hop question-answering harness.
Artificial intelligence (AI) is a branch of computer science that enables machines to perform tasks associated with human intelligence, including learning, reasoning, pattern recognition and decision-making. Modern AI generally learns patterns from large datasets, tunes internal weights and parameters during training, evaluates outputs on held-out data, and applies the resulting model during inference to classify, predict, generate content or trigger actions. Most organizations fine-tune existing foundation models rather than train from scratch, while output quality remains dependent on the completeness, bias and quality of training data. The page separates reactive machines and limited memory from theoretical theory of mind and self-aware systems, and distinguishes today’s narrow AI from theoretical general AI and superintelligence. It also describes generative AI, common applications, risks including hallucinations, bias, privacy and security gaps, and governance, concluding that practical adoption depends on real problems, trusted data and responsible oversight.
Parameter-efficient fine-tuning (PEFT) reduces the memory required to adapt open models, enabling fine-tuning with a fraction of the memory and, in some cases, quantized models. LoRA is dominant because it freezes base weights and trains a small set of added parameters, but its popularity may also reflect ecosystem momentum and uneven comparisons. The post presents shared-condition benchmarks for language-model math reasoning and image generation, using identical models, datasets, code, hardware, and measures including test performance, VRAM, forgetting or drift, runtime, and checkpoint size. The results show that LoRA performs well but is not consistently best: other techniques can outperform it on one or more axes, especially in the image-generation benchmark. The recommended choice is therefore to compare techniques against a use case’s needs and constraints, including quantization support, layer coverage, adapter merging, and downstream serving compatibility.
Final-answer checks can show whether an agent completed a task, but they miss the turns, tokens, seconds, errors, and command path required. The proposed tool-specific benchmark measures that work on transformers, comparing models, revisions, tasks, and three access tiers: bare installation, source clone, and a packaged Skill containing CLI documentation and examples. It uses deterministic tasks with exact matching, launches one Hugging Face Job per model–revision–task combination on identical hardware, and records traces in a Hugging Face Bucket. Reports include match rate, median new, cached, and generated tokens, time, error rate, and marker adoption, while traces expose the agent's commands. On transformers, the CLI plus Skill helped large open models but caused Qwen3-14B to misinterpret documentation as an executable tool, dropping classify-sentiment match from 100% to 0% in the cited comparison.
Payment fraud detection is presented as a real-time practice for identifying and blocking unauthorized transactions amid evolving card-not-present, account takeover, authorised push payment, friendly, and card-testing attacks. Modern systems combine predefined rules, machine learning, behavioral analytics, device fingerprints, geolocation, transaction history, and velocity signals to calculate risk scores within milliseconds, declining or routing high-risk payments to review. Rules are fast and interpretable but require continual tuning, while machine-learning and behavioral models can identify complex or user-specific patterns that static thresholds miss. The source also emphasizes prevention controls such as MFA, tokenization, 3D Secure, and device binding, alongside unified real-time data infrastructure that supports cross-channel signals, continuous retraining, and deployment into payment authorization flows.
Strands Robots integrates Hugging Face’s LeRobot stack with Strands Agents to connect dataset recording, policy execution, simulation, hardware deployment, and fleet coordination in one agent loop. Its integration keeps LeRobot responsible for hardware drivers, calibration, recording, and dataset handling, while Strands supplies AgentTools, a shared Robot abstraction, policy providers, and a Zenoh peer mesh; simulation and hardware use the same LeRobotDataset format. The walkthrough records demonstrations in MuJoCo, optionally pushes them to the Hub, runs policies through a common interface, and switches from simulation to a physical SO-101 with mode="real". The companion script and notebook run the default simulation workflow on a laptop without hardware, a GPU, or Hugging Face credentials, although the default Mock policy produces structurally valid but behaviorally unhelpful random demonstrations.
GLM-5.2 is presented as a flagship model for long-horizon tasks, extending GLM-5.1’s capability to a solid 1M-token context for sustained engineering work. It was trained extensively on 1M-context coding-agent scenarios involving large-scale implementation, automated research, performance optimization, and complex debugging, and adds effort-level controls to balance capability, latency, and computational cost. Its architecture uses IndexShare, sharing one lightweight indexer across every four transformer layers and reducing per-token FLOPs by 2.9× at 1M context; MTP changes for speculative decoding increase acceptance length by up to 20%. On reported long-horizon and standard coding benchmarks, GLM-5.2 is described as the highest-ranked open-source model: it trails Opus 4.8 by 1% on FrontierSWE, ranks second only to Opus 4.8 on PostTrainBench, and scores 81.0 on Terminal-Bench 2.1 versus GLM-5.1’s 63.5.
Building AI experiences for augmented reality and wearable devices requires bridging hardware with live media streams, multimodal models, enterprise tools, and runtime infrastructure. NVIDIA XR AI provides an open-source, modular framework connecting extended reality headsets and smart glasses to GPU-accelerated services across cloud, edge, and workstations. In this architecture, camera frames and microphone audio ingest into an XR Media Hub that routes data while keeping raw video pixels in shared memory to minimize overhead. The ecosystem uses NVIDIA Cosmos models for vision-language grounding, NVIDIA Nemotron models for reasoning and tool invocation, and the Model Context Protocol for enterprise integrations. Optional agent orchestration via NVIDIA NeMo Agent Toolkit and spatial streaming through NVIDIA CloudXR support complex workflows across healthcare and manufacturing.
Production financial intelligence systems frequently rely on hand-engineered tabular features and rule sets that are brittle, expensive to maintain, and blind to historical sequential patterns. NVIDIA demonstrates an accelerated reference pipeline that replaces standard BPE tokenization with a GPU-based domain tokenizer, converting raw transactions into semantic tokens with an 8,192-token context window. Using the NeMo AutoModel library, a compact 29-million parameter decoder-only transformer is pretrained from scratch on unlabeled transaction sequences using causal language modeling. Learned sequence embeddings are extracted, compressed using PCA, and concatenated with raw tabular features to train a downstream GPU-accelerated XGBoost fraud detection model. On the IBM TabFormer benchmark dataset, this combined approach achieves a 41.76% lift in Average Precision over the baseline model relying solely on raw tabular features.
NVIDIA submitted results across all MLPerf Training v6.0 benchmarks, setting performance records on workloads including the DeepSeek-V3 and GPT-OSS-20B Mixture of Experts models. To scale training up to 8,192 Blackwell GPUs, NVIDIA combined hardware cluster designs with scale-out networking via Spectrum-X Ethernet and Quantum InfiniBand. The engineering team eliminated CPU-GPU synchronization bottlenecks in token-dropless MoEs by implementing full-iteration CUDA graphs with synchronization-free operators and paged stashing. Further software optimizations included CuTe DSL kernel fusions, an MXFP8 attention block in the Transformer Engine, and transitioning router elementwise math to FP32. Across these benchmarks, the GB300 NVL72 platform achieved the fastest time to train at scale and the highest normalized per-accelerator performance.
Game developers integrating on-device artificial intelligence companions face challenges such as conversational latency, state synchronization, and loop prevention. NVIDIA announced the open-source ACE Game Agent SDK alongside a suite of Unreal Engine 5 plugins to streamline native, hardware-accelerated character pipelines. The SDK offers Agent, Chat, and RAG APIs, enabling characters to execute multi-step tool-assisted reasoning and query game databases. Unreal Engine 5 plugins introduce local runtime models including nemo-conformer-ctc-120m for speech recognition, Qwen 3.5 4B for language generation, and Chatterbox Turbo 350M for speech synthesis. Additional tooling expands to motion generation via Animotive Kimodo and rendering updates with the DLSS 4.5 Unreal Engine plugin.
Accelerating transformer training requires understanding how low-precision formats such as FP8 and NVFP4 impact specific general matrix multiplication (GEMM) workloads. Transformer configurations do not explicitly reveal these shapes, making microbenchmarks necessary before committing to full training runs. NVIDIA Transformer Engine enables quantization and kernel dispatch across precisions, which can be evaluated in realistic autocast mode or kernel-only prequantized mode. Profiling the ESM2-15B model on NVIDIA B300 GPUs demonstrated that NVFP4 achieved a 1.79x blended forward propagation speedup over MXFP8 and up to 4.01x over BF16 in prequantized execution. Although large GEMM dimensions successfully overcome quantization overheads, dynamic scaling, Hadamard transforms, and kernel selection asymmetries moderate real-world gains.
Customer segmentation divides an existing customer base into smaller groups based on shared demographic, geographic, psychographic, behavioral, firmographic or value-based characteristics. Unlike market segmentation, it uses first-party data about existing customers, and a customer can belong to multiple segments at once. The guide distinguishes rule-based, survey-based, RFM, k-means, decision-tree and AI/ML-driven methods, with choices depending on data maturity and business goals. Effective implementation defines an objective, audits and unifies sources, resolves duplicate identities, selects a method, validates segments and measures outcomes. It also describes Databricks' CustomerLake capabilities, including governed Customer 360, agentic identity resolution, natural-language segmentation through Genie and bidirectional activation connectors; stated benefits include improved retention, conversion, customer lifetime value and marketing efficiency.
Wayfair’s Applied Research team used Cursor to automate and parallelize experiments for a product-tag validation model that checked attributes against images, descriptions, and reviews. The model was accurate but too expensive to run across the catalog. In December 2025, five researchers standardized the dataset and evaluation benchmark, then built and tested 110 distinct variants in four days; the winning architecture reduced inference costs by 94%, improved precision, and became the production baseline. In March 2026, the team benchmarked against that baseline, ran 140+ new experiments, and added genetic-algorithm searches, achieving another 90% cost reduction. The workflow used 20+ parallel agents, desktop and CLI surfaces, and cloud agents, leaving researchers to formulate hypotheses, interpret results, and choose candidates while Cursor implemented, ran, and measured variants.
Sidekick’s customer-segmentation skill was fine-tuned on tens of thousands of successful, de-identified production queries, so it generated zero-result queries instead of refusing impossible requests such as finding customers by occupation. A small Toloka dataset of roughly 600 standard queries and 602 refusal annotations became a seed for four frontier LLM judges, which were calibrated against human-labeled examples and required to agree on both verdict and reasoning. The pipeline used mutually exclusive categories to resolve conflicting labels, filtered disagreements for human review, and fed accepted examples from later production traffic back into fine-tuning. With automated curation, the segmentation pass rate rose from 0.762 to 0.798 versus naive merging, while manual validation measured 86.3% refusal accuracy and a 4.6% false-positive rate. The resulting flywheel is being applied to other Sidekick skill models and data-quality problems.
Sentry upgraded its AI-driven issue grouping system to better prevent duplicate issues without merging distinct application errors. The team trained lightonai/modernbert-embed-large using Matryoshka Representation Learning on hundreds of thousands of stacktrace pairs labeled by Claude Sonnet 4.5. To optimize the high-throughput ingestion pipeline, embeddings were truncated from 768 to 64 dimensions, combined with bfloat16 precision, PyTorch SDPA, and CUDA graph compilation. A phased live rollout used threshold-gated index backfilling and automatic fallback to prevent issue spikes during migration. In production, the v2 model cuts the overgrouping rate from 8% to 4%, increases prevented duplicate issues to 70%, and delivers 6x faster inference.
Databricks announces its inaugural Student Fellows cohort, selected from more than 5,000 applications submitted by students at hundreds of universities and from dozens of countries. The program targets students who contribute on campus, apply data and AI in practice, and intend to pursue careers in the field, with fellows acting as bridges between academic theory and the real-world scale of the Databricks platform. Five students are profiled, with experience spanning large-scale AI systems, ETL pipelines, computer vision, machine learning platforms, robotics, data modeling, and local retrieval-augmented generation architectures. During the coming academic year, the cohort will receive training from Databricks experts and hands-on experience solving complex data challenges, while building a launchpad for future internship opportunities. The program invites applications for its next cohort in Fall 2026 and points readers to Databricks Free Edition.
The best engineering reads, once a week.
