The Archive

iWorld-Bench is a 330k-clip dataset and benchmark for training interactive world models on perception, reasoning, and physical interaction capabilities.

Study finds LMs can iteratively refine conceptual definitions through counterexample generation, but accept invalid counterexamples at 2× the human acceptance rate.

Open-World Sound Event Detection paradigm extends closed-world audio classifiers to detect unknown events and incrementally learn from them.

Magic-informed quantum architecture search uses GNN-guided Monte Carlo Tree Search to control quantum resource utilization in circuit design.

PHALAR contrastive framework for stem retrieval uses learned spectral pooling and phase-equivariant biases, achieving 70% relative accuracy gain with fewer parameters.

Randomized algorithm for PAC policy identification in MDPs combines posterior sampling with online learning for asymptotically optimal sample complexity.

Theoretical proof that dyadic refinement iterates on R² admit exact ReLU realizations with fixed width and O(n) depth for piecewise-linear functions.

RCT of 356 clinicians shows atomic fact-checking (decomposing LLM recommendations into verifiable claims) increases trust from 27% to 67% vs. traditional explainability methods.

Task vector arithmetic on BEATs encoders composes 661-species bioacoustic classifier without data sharing; task vectors near-orthogonal, geometry aligns with acoustic niche hypothesis.

Google releases Gemma 4 multi-token prediction drafters in 4 quantized sizes for local deployment.

Google partners with XPRIZE and Range Media on $3.5M Future Vision film competition.

Anthropic releases ten Cowork and Claude Code plugins plus Microsoft 365 integrations and MCP app for financial services.

The automotive cockpit is undergoing a fundamental shift from rule-based interfaces to agentic, multimodal AI systems capable of reasoning, planning, and... The automotive cockpit is undergoing a fundamental shift from rule-based interfaces to agentic, multimodal AI systems capable of reasoning, planning, and acting. In most vehicles on the road today, in-vehicle assistants still rely on fixed command-response patterns: interpret a phrase, trigger an action, reset. While effective for well-defined tasks, this approach doesn’t scale to modern… Source

Framework shows popular activation steering methods misalign with prompt steering mechanics; proposes distilling prompt behavior into interpretable models to close performance gap.

CC-OCR V2 benchmark for real-world enterprise document OCR with LMMs; addresses gap between lab tasks and practical heterogeneous acquisition conditions.

Gauge-invariant GNN architecture for Abelian lattice gauge theories using Wilson loop representations; application to condensed matter and quantum systems.

Argues frontier AI failures in open-ended tasks (scientific assistance, agents, personalization) stem from objective ambiguity rather than capability gaps; proposes contextual multi-objective optimization.

Reddit post on using Qwen3.6 with pi.dev harness and agent tooling for local coding and admin tasks.

Generative AI’s explosive first chapter was defined by humans sending requests and models responding. The agentic chapter is different. Agents don't... Generative AI’s explosive first chapter was defined by humans sending requests and models responding. The agentic chapter is different. Agents don’t follow a pre-determined sequence of actions. They call tools, spawn sub-agents with different tasks and models, retain information in memory, manage their own context window, and decide for themselves when they’re finished. In doing so… Source

Process-aware pipeline for continuous predictive monitoring of clinical pathways using prefix-based representations on COVID-19 ICU admission prediction.

Link: x.com

Google demonstrates 3X LLM inference speedup on TPUs using diffusion-style speculative decoding technique.

PayPal is pitching an AI-led turnaround, tying automation and restructuring to $1.5B in savings as it cuts jobs and works to modernize its tech stack.

User reports 'Stream ended without a final message' error in Claude Design, a feature for sketching animations.

Proposes improved empirical fixation density estimation methods beyond fixed-bandwidth Gaussian KDE for saliency benchmarking and per-image model evaluation.

QKVShare framework for quantized KV-cache handoff between multi-agent LLMs on edge devices; token-level mixed-precision allocation reduces memory vs. full-precision transfer.

Dual-Embodiment Companion Framework extends AI capabilities to personal physical objects (plush toys); formative study derives design principles for emotional continuity.

ProgramBench: 200-task evaluation showing agents struggle to rebuild large binaries from scratch without cheating vulnerabilities.

DMGD proposes training-free dataset distillation using diffusion models with semantic-distribution matching guidance.

Study compares 2D spatiotemporal convolutions vs. concatenated 1D convolutions for EEG signal classification with CNNs.