AItira Keyword Portal

#AI 트렌드

AI 트렌드 키워드와 연결된 AI 도구, 교육·행사, 보드 인사이트, 블로그, 뉴스레터 콘텐츠를 AItira에서 한 번에 확인하세요.

전체 362개Finder 0개Calendar 0개Board 362개Blog 0개Newsletter 0개

관련 AI 인사이트 362

Understanding AI from the Inside: The State of Mechanistic Interpretability for AI SafetyNeural networks are likely to be among the most consequential artefacts in human history. As these systems bec…Label-Wording Augmentation in a Typed-Decision Encoder: Gains on Unfamiliar Words, Not on Abbreviations or CodesTyped-decision models score options that the caller supplies at request time, so their accuracy depends on lab…ReproBridge: Building Reproducibility into ML Research ArtifactsReproducing machine learning experiments requires reconstructing procedures from information scattered across…Towards interpreting any residual stream direction by scaling activation inversionTo interpret a direction in a language model's residual stream, we can look for text that strongly activates i…Same Behavior, Different Mechanism: What Finite Experiments Can Certify About Integrated InformationInterpretability, AI auditing and assessments of machine consciousness increasingly rest on mechanism-level qu…XStack: A Benchmark for ML Repo MigrationML code should run where researchers choose, not where it was first written. Today, machine learning code main…Scientist World: Grow a Research World for AI-ScientistsAI-Scientists arrived, but a world for them to grow in did not. Human scientists take theirs for granted $-$ c…SwarmEvolve: Latent Agent Swarms for Evolutionary Discovery and OptimizationLLM-guided evolutionary search uses AI agents to improve programs for mathematical and computational problems…Flipsnack grows product signups with conversion optimizationFlipsnack focused its ChatGPT Ads campaigns on getting people to try its product, growing registrations while…Merge builds product and business tools with OpenAI technologyMerge uses GPT-6 Astra, ChatGPT, and Codex to review product code and support marketing. Its teams also use AI…Cult.fit builds operational muscle with CodexCult.fit is using Codex to help teams across functions build tools, move faster, and change how work gets done…Australian universities are creating AI-native graduatesAustralian universities are expanding access to ChatGPT and bringing AI into teaching and learning to prepare…Measuring how AI agents handle everyday wet-lab scienceOpenAI introduces BioBasicsBench, a 100-task benchmark measuring how well AI agents interpret raw experimental…Partnering with Hacktron to find vulnerabilities in OpenAI’s systemsOpenAI is partnering with Hacktron to find vulnerabilities, test connections with third-party services, and st…CounterFactual Retrieval for Evidence-Aware Individual Animal IdentificationAutomated identification of individual animals is moving from small curated studies toward shared benchmarks a…On Turn-Level Best-of-N Sampling for Scaling LLM AgentsIn this work, we evaluate a simple inference-time scaling approach based on \textit{turn-level best-of-N sampl…Learning to Decode Linear Codes with Generalized Belief Propagation NetworkError correcting codes are a fundamental component of communication systems, enabling reliable data transmissi…Dull, Dirty, Dangerous? Working Conditions and the Limits of Occupational AI ExposureAutomation is often justified as relief from dull, dirty, and dangerous work. I examine how these conditions r…Occlusion for Free: Detecting Occlusion from the Disagreement Between Two Monocular Distance CuesKnowing that a pedestrian is occluded is useful to a robot even when the detector cannot recover the hidden pi…Perceptive Humanoid Motion from SimulationHumanoid robots have the potential to automate a wide range of physical tasks. Progress toward generalist huma…FrontierPhysics: Evaluating Agents for End-to-End Physics ResearchAI agents are playing an increasingly important role in physics research. However, existing physics benchmarks…BEAKER: An Expert-Curated Benchmark for Embodied Brains in Self-Driving Chemical LaboratoriesSelf-Driving Chemical Laboratories (SDCLs) are moving chemical experimentation toward embodied automation, whe…DevRev delivers autonomous self-service with GPT-6.1 SolDevRev uses GPT-6.1 Sol to power Computer and Operator, balancing reasoning, speed, and cost for enterprise se…Introducing Images 2.6Images 2.6 brings faster image generation, GIF creation, more precise editing, and new API controls across Cha…A thousand fewer paths to rootOpenAI built an autonomous pipeline that found, validated, and generated patches for 948 local privilege-escal…Setting a new standard for responsible AI infrastructureOpenAI introduces its National Community Compact and Compute Campus Policy Blueprint to protect local resource…GPT‑6 Cyber: Advancing the frontier of aligned cybersecurity agentsIntroducing GPT‑6 Cyber, OpenAI’s cybersecurity model for finding exploitable vulnerabilities and developing f…A simple adaptive accelerated optimistic dual averaging method for unconstrained convex optimization in $p$-norm geometryEfficient preconditioning plays an important role in large-scale first-order optimization. In this paper, we p…MedPhysBench: A Benchmark for Evaluating Language Models on Medical Physics Calculations and Error DetectionMedical physicists turn a prescription into a delivered dose and check one another's calculations, because one…Weight-space programming for auditing clinical decision networksMachine learning models in radiotherapy are commissioned by testing them on cases with known answers, which ma…Adversarial Speculative Reasoning Improves Long Horizon AgentsSpeculative decoding has been treated purely as a systems optimization that speeds up large language model dec…Suppressing Misaligned Behaviors in Language Models, DurablyUndesired model behaviors can emerge or re-emerge at any stage of training. Once observed, how effectively can…Aligning Language Models with General Preferences: An Empirical Study of Direct Nash OptimizationDirect Nash Optimization (DNO) aligns a language model to a general preference function by approximating the N…From Autonomous Agents to Cognitive Symbionts: Toward AI-Augmented Human IntelligenceFor a decade, progress in artificial intelligence has been measured by how much a machine can do on its own. L…DreamX: Video Diffusion Model is a Universal Latent Action LearnerLearning a universal action representation from video requires modeling dynamics that are consistent across di…Agentic cheminformatics with Google AntigravityLearn how Google Antigravity and our Cheminformatics Skills accelerate drug discovery workflows, search compou…