Skip to main content
← All Tags

Machine Learning

280 articles in this category (Page 2 of 12)

AI NewsMachine LearningArtificial Intelligence

Meta AI Open-Sources NeuralBench: A Standardized Benchmark for EEG Foundation Models

Meta AI's NeuralBench-EEG v1.0 standardizes NeuroAI evaluation across 36 tasks and 94 datasets, revealing that 150K-parameter models often rival 157M-parameter foundation models.

Read more
AI NewsMachine LearningCloud Computing

Mastering Gemma 4 Fine-Tuning: Fixes for ClippableLinear and Multimodal Masking

Gemma 4 fine-tuning requires specific 'all-linear' LoRA targeting and backward-search masking to achieve 94.2% accuracy on multimodal tasks.

Read more
AI NewsArtificial IntelligenceMachine Learning

Secure Non-Deterministic AI Agents with Statistical Guardrails

Secure AI agents using cosine distance z-scores and Shannon entropy to detect semantic drift and low-confidence hallucinations in real-time.

Read more
AI NewsMachine LearningArtificial Intelligence

How to Build an End-to-End Production Grade Machine Learning Pipeline with ZenML

Learn to build production-grade ML pipelines using ZenML with custom materializers, metadata tracking, and fan-out hyperparameter optimization.

Read more
AI NewsAI InfrastructureMachine Learning

Zyphra's TSP Strategy Achieves 2.6x Throughput for Large-Scale AI Training

Zyphra introduces Tensor and Sequence Parallelism (TSP), a hardware-aware strategy delivering 2.6x throughput over TP+SP baselines using 1,024 AMD MI300X GPUs.

Read more
AI NewsData ScienceMachine Learning

Correcting Survey Bias with Meta's balance Library: A Technical Guide

Learn to eliminate sampling bias using Meta’s balance library, featuring IPW and CBPS methods to restore survey accuracy.

Read more
AI NewsLarge Language ModelMachine Learning

TaskTrove: A Technical Workflow for Streaming Parsing and Verifier Detection

Efficiently stream and parse the multi-gigabyte TaskTrove dataset to detect RL-ready verifier signals using real-time binary decoding and automated visualization.

Read more
AI NewsAgentic AIMachine Learning

Building Multi-Agent AI Workflows for Advanced Systems Biology Simulations

Develop a multi-agent AI pipeline using GPT-4o-mini to model gene networks, predict protein interactions, and optimize metabolic flux with unified LLM-driven synthesis.

Read more
AI NewsMachine LearningEngineering

Calculating Local LLM VRAM Requirements to Prevent GPU Out-of-Memory Errors

Master the mathematics of LLM VRAM consumption, from the 2-byte-per-parameter baseline to KV cache overhead and 4-bit quantization savings.

Read more
AI NewsAgentic AIMachine Learning

Meta Autodata: Agentic Framework for High-Quality Training Data Creation

Meta AI introduces Autodata, an agentic framework that enables autonomous data creation, increasing performance gaps between model solvers from 1.9% to 34%.

Read more
AI NewsAI InfrastructureMachine Learning

Qwen-Scope: Open-Source Sparse AutoEncoders for LLM Interpretability and Steering

Qwen AI releases Qwen-Scope, an open-source suite of 14 Sparse AutoEncoders (SAEs) for Qwen3/3.5 models, enabling inference-time steering and benchmark analysis without model runs.

Read more
AI NewsMachine LearningAI

Transformer Output Selection: Softmax and Fully Connected Layer Integration

Learn how Transformer decoders transform terminal residual values into vocabulary-mapped outputs using fully connected layers and softmax for token prediction.

Read more
AI NewsMachine LearningSoftware Engineering

Inside OpenAI's Parameter Golf: Training High-Performance LLMs in 10 Minutes

OpenAI's Parameter Golf challenge requires training a 16MB language model in 10 minutes, with top developers reaching 1.0810 bits-per-byte.

Read more
AI NewsArtificial IntelligenceMachine Learning

AI-Driven ML: Automating Time-Series Forecasting with Anton

MindsDB introduces Anton, an open-source AI agent that automates the end-to-end ML lifecycle, achieving a 14.6% MAPE on demand forecasting within minutes.

Read more
AI NewsAI InfrastructureMachine Learning

FlashQLA: High-Performance Linear Attention Library for NVIDIA Hopper GPUs

The Qwen Team has released FlashQLA, a linear attention kernel library achieving up to 3x speedup on NVIDIA Hopper GPUs for Gated Delta Network architectures.

Read more
AI NewsMachine LearningSoftware Engineering

OpenAI Privacy Filter: Building a Production PII Redaction Pipeline

Learn to implement a production-grade PII detection pipeline using the OpenAI Privacy Filter to automatically identify and redact sensitive data like API keys and personal addresses.

Read more
AI NewsComputer VisionMachine Learning

Best of WACV 2026: Advances in Zero-Shot Sampling and OOD Detection

Join Voxel51 on April 30 for the Best of WACV 2026 virtual event featuring four technical talks on subspace sampling and MLLM robustness.

Read more
AI NewsAgentic AIMachine Learning

Optimizing Long-Term Memory Retrieval with Reinforcement Learning for LLM Agents

Build a PPO-trained RL agent that optimizes long-term memory retrieval for LLMs, outperforming standard cosine similarity in complex QA tasks.

Read more
AI NewsMachine LearningSoftware Engineering

RMS Normalisation and Residual Connections: Stabilizing Deep Neural Networks

Stabilize deep networks by preventing activation drift and vanishing gradients using RMSNorm and residual connections for efficient training.

Read more
AI NewsMachine LearningComputer Vision

Meta AI Sapiens2: Scaling Human-Centric Vision Models to 5B Parameters and 4K Resolution

Meta AI's Sapiens2 scales to 5B parameters and 1B images, achieving 82.3 mAP in pose estimation and 82.5 mIoU in segmentation across 1K and 4K resolutions.

Read more
AI NewsLarge Language ModelMachine Learning

Talkie-1930: A 13B Vintage LLM Trained Exclusively on Pre-1931 Data

Researchers released Talkie-1930, a 13B parameter open-weight LLM trained on 260 billion tokens of pre-1931 text to eliminate benchmark contamination and research historical reasoning.

Read more
AI NewsArtificial IntelligenceMachine Learning

Optimizing CJK Text Wrapping with BudouX Machine Learning Parsers

Learn to implement BudouX for phrase-aware line breaking in Japanese, Chinese, and Thai, utilizing lightweight ML models to process text at speeds exceeding 1,000k chars/sec.

Read more
AI NewsMachine LearningWeb Development

Local Browser-Based AI: Running Neural Networks for Audio Stem Separation

Stem separation moves to the edge as Demucs v4 runs in a browser tab via ONNX and WASM, processing a 4-minute song locally in 3-5 minutes.

Read more
AI NewsMachine LearningSoftware Engineering

Implementing Microsoft’s OpenMementos: Trace Analysis and Context Compression for LLMs

Implement Microsoft’s OpenMementos dataset to achieve ~6× token compression in reasoning traces for efficient LLM fine-tuning and inference.

Read more