Topics
Browse by topic
Browse every AI research topic on Research Papers — language models, diffusion, agents, world models, embodied AI, and more.
Language Models · 59 Models trained to understand, generate, and transform natural language at scale. LLM Reasoning · 51 Eliciting and improving step-by-step reasoning in large language models. AI Agents · 48 LLM-driven systems that plan, act, use tools, and carry skills across tasks. Multimodal Models · 44 Foundation models that combine language with images, audio, video, or other signals. Efficient AI · 40 Algorithms and systems that reduce memory, compute, or latency for large models. Diffusion Models · 34 Generative models that synthesize data through iterative denoising. Vision Foundation Models · 23 Large visual representation models that transfer across recognition, localization, and perception tasks. Fine-Tuning & Adaptation · 19 Adapting pretrained models to new tasks cheaply, including parameter-efficient methods like LoRA. Text-to-Image · 19 Models that generate or edit images from natural-language prompts. Transformers · 17 Attention-based architectures that became the backbone of modern language and multimodal models. Alignment · 15 Methods for steering models toward preferred, safer, or more useful behavior. Open Models · 15 Open-weight model releases and the training recipes behind them. Reinforcement Learning · 15 Training language models and agents from reward — RLHF, RLVR, GRPO, and verifiable-reward methods that drive reasoning gains. Robotics · 15 Learning and control for physical robots. Retrieval-Augmented Generation · 14 Grounding language model outputs in retrieved documents to improve factuality and freshness. World Models · 14 Generative models that simulate consistent, controllable environments over time. AI for Science · 13 Machine learning applied to scientific discovery — biology, chemistry, physics, and materials, from protein structure to new materials. Long Context · 13 Models and evaluations for reasoning over very large text, audio, video, or code contexts. Vision-Language-Action · 12 Models that map perception and language directly to robot actions. Agent Memory · 10 How AI agents store, retrieve, and update long-term memory across tasks and sessions — beyond the context window. Video Generation · 9 Models that synthesize video from text or other conditions, including streaming and autoregressive diffusion approaches. Code Generation · 8 Models and systems that synthesize, complete, or reason about programs. Sequence Modeling · 8 Architectures for modeling long ordered data such as text, audio, code, and genomics. Mixture of Experts · 7 Sparsely activating subsets of parameters so model capacity grows without proportional compute. Self-Supervised Learning · 7 Training methods that learn useful representations from data without task-specific labels. Theorem Proving · 7 Neural, symbolic, and hybrid systems for mathematical proof search. Biomolecular Modeling · 6 AI systems for protein structures, molecular interactions, and computational biology. Brain Decoding · 6 Using machine learning to read, model, and causally probe how the brain represents perception. Diffusion Language Models · 6 Text generation by iterative denoising instead of left-to-right decoding — parallel, non-autoregressive language models. Segmentation · 6 Promptable and automatic systems for separating objects in images and videos. Small Language Models · 6 Compact, on-device, and edge-deployable models — strong capability per parameter for local and low-cost inference. Speech Synthesis · 6 Text-to-speech and voice generation models, including zero-shot, expressive, and dialogue synthesis. Text Embeddings · 6 Methods for turning text into dense vectors for retrieval, similarity, and search, including using LLMs as encoders. Interpretability · 5 Reverse-engineering what neural networks compute inside — features, circuits, and the mechanisms behind model behavior. Speech Recognition · 5 Models for transcribing, translating, and understanding spoken audio.