New Model Releases A
Showing 181–210 of 312
-
Skillful forecasting of offshore winds from satellite scatterometer constellations
-
MindForge: Teaching Small Language Models Whole-Life-Cycle Software Engineering via Source-Free Program Synthesis
-
DLAM: Distributional Latent Actions with Temporal Constraints
-
AgentMap: Joint Equivalence and Subsumption Discovery for Ontology Matching
-
Hierarchical Spatio-Temporal Transformer for Coherent Emergency Department Forecasting
-
InferScale: GPU-Native KV Injection for Personalized LLM Serving
-
SciFigQual-Bench: A Benchmark for Scientific Figure Quality Assessment with Full-Manuscript Context
-
MemSecBench: Tracking Agent Memory Poisoning from Persistence to Consequence and Repair
-
We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative controlGoogle DeepMind brings Lyria 3.5 music model to Google Flow MusicGoogle DeepMind announced that Lyria 3.5, its AI music-generation model, is launching in Google Flow Music, with reported advances in musicality, lyrics, vocals, and creative control. Details beyond the title are unconfirmed as the source excerpt was empty.
-
Equilibrium Training of Energy-Based Models with Parallel Trajectory Tempering
-
Single-Beat Cuffless Blood Pressure Estimation Using Ear-PPG and ECG with a Lightweight Hybrid Learning Framework
-
SciFigAlign: Scoring Scientific Figures by Fine-tuned Alignment of Visuals with Manuscript Evidence
-
PIKS: Universal Physics-Informed Kernel Methods
-
HoF-Bench: Rediscovering Real AI-Discovered CVEs Without Frontier Models
-
BayesAME: Bayesian Active Model Evaluation
-
Evaluating Regional Bias in LLMs From Abstract Stereotype to Concrete Social Decision-Making
-
Sakana AI防衛・インテリジェンスチーム、「DIVER OSINT CTF 2026」で5位入賞 Fuguを活用したOSINTエージェントの可能性Sakana AI's defense team places 5th at DIVER OSINT CTF 2026Sakana AI's defense and intelligence team placed fifth at the DIVER OSINT CTF 2026 competition, using its Fugu tool to power an OSINT agent. The result highlights the potential of AI agents for open-source intelligence and information-analysis tasks.
-
How enabling two settings tripled our scores on the ARC-AGI-3 benchmarkOpenAI triples ARC-AGI-3 scores by enabling two API settingsOpenAI reported that enabling two API settings tripled GPT-5.6's scores on the ARC-AGI-3 benchmark. By retaining reasoning across calls and tuning configuration, the setup improved both accuracy and efficiency, which the post breaks down in detail.
-
AgentSnare: Learning to Delay, Divert, and Defuse Autonomous Penetration Agents
-
OptimismBench: Forecasting Bias and the Alignment Effect in Language Model Judgment
-
TREK: A Travel Reasoning and Evaluation Kit for LLM Agents in Complex Trip Planning
-
Generation or Judgement? A Paradigm Perspective on LLM-Based Emotion-Cause Pair Extraction in Conversation
-
Feature Bagging Provides Stability
-
Credit Cards, Confusion, Computation, and Consequences: What Can We Uncover About Language Model Reasoning?
-
Progressive Multimodal Alignment for Continual Instruction Tuning
-
Belief-Guided Decision Making with Uncertainty Gating in the Game of Go
-
Surrogate assisted diversity estimation in neural ensemble search
-
Siobahn Day Grady Wants Everyone to Be AI LiterateNCCU's Grady launches first HBCU AI institute, pushes AI literacyIEEE Spectrum profiles Siobahn Day Grady, an NCCU associate professor who in January 2025 launched IAIER, the first AI research institute at a historically Black college or university (HBCU). She aims to broaden AI literacy and narrow the funding and industry-access gaps spread unevenly across higher education. The excerpt is truncated, so IAIER's specific activities are unconfirmed.
-
Defending Against Backdoor Attacks via Alignment Checking in Model-Contrastive Federated Learning
-
Same Evidence, Different Target: Decoding How Diagnostic Evidence Bears on Causal Questions from Language-Model States