📚 arXiv最新AI论文速览速学

菜单

📄

提交新论文

AI论文阅读
🔍 搜索与筛选
大类标签
所有标签
reinforcement learning 52 diffusion models 26 benchmark evaluation 21 mathematical reasoning 20 policy optimization 16 diffusion transformer 13 multi-agent systems 12 tool usage 12 automated evaluation 11 chain-of-thought 11 multimodal llm 11 retrieval-augmented generation 11 code generation 10 efficient inference 10 knowledge distillation 9 attention mechanisms 8 benchmarking 8 embodied ai 8 evaluation framework 8 instruction tuning 8 preference learning 8 synthetic data generation 8 text-to-image generation 8 vision-language models 8 vision-language-action 8 mixture-of-experts 7 multimodal evaluation 7 multimodal reasoning 7 robot manipulation 7 self-supervised learning 7 video generation 7 web agents 7 efficiency optimization 6 gaussian splatting 6 image generation 6 reward modeling 6 benchmark dataset 5 computational efficiency 5 contrastive learning 5 data generation 5 data synthesis 5 diffusion language models 5 evaluation benchmark 5 gui agents 5 image editing 5 multi-hop reasoning 5 multimodal llms 5 preference optimization 5 reasoning models 5 robot control 5 scaling laws 5 transformer architecture 5 visual reasoning 5 3d generation 4 3d reconstruction 4 attention mechanism 4 benchmark 4 curriculum learning 4 data efficiency 4 domain adaptation 4 low-resource languages 4 model scaling 4 multi-agent collaboration 4 multi-agent framework 4 multi-step reasoning 4 physical reasoning 4 progressive training 4 synthetic data 4 text generation 4 video diffusion models 4 visual language model 4 visual question answering 4 zero-shot learning 4 adaptive reasoning 3 adversarial attacks 3 autonomous driving 3 character animation 3 code execution 3 content moderation 3 credit assignment 3 data augmentation 3 data curation 3 distributed training 3 efficient training 3 explainable ai 3 hallucination detection 3 inference acceleration 3 inference optimization 3 instruction following 3 language modeling 3 large language models 3 llm evaluation 3 llm-as-judge 3 long video generation 3 long-context reasoning 3 long-horizon tasks 3 low-rank adaptation 3 memory optimization 3 mixture of experts 3 model alignment 3 model context protocol 3 motion generation 3 multi-turn reasoning 3 multimodal benchmark 3 multimodal foundation model 3 multimodal fusion 3 out-of-distribution generalization 3 parameter efficiency 3 parameter efficient fine-tuning 3 positional encoding 3 pseudo-labeling 3 reasoning benchmark 3 reasoning efficiency 3 reasoning tasks 3 reflective reasoning 3 research agents 3 rl training 3 safety alignment 3 safety evaluation 3 scientific ai 3 self-evolving agents 3 sparse attention 3 sparse rewards 3 state space models 3 text-to-image 3 text-to-video 3 tool integration 3 transformer optimization 3 unified architecture 3 verifiable rewards 3 video question answering 3 video understanding 3 virtual try-on 3 vision-language model 3 3d editing 2 3d gaussian splatting 2 4d generation 2 adaptive retrieval 2 adaptive sampling 2 adversarial robustness 2 agent architecture 2 agent framework 2 agentic reasoning 2 ai safety 2 alignment drift 2 attention masking 2 audio-driven animation 2 automated assessment 2 automated reasoning 2 autoregressive generation 2 autoregressive model 2 autoregressive models 2 benchmark design 2 benchmark framework 2 browser automation 2 chain-of-thought reasoning 2 clinical evaluation 2 cognitive modeling 2 context management 2 continuous pretraining 2 controllable generation 2 creative writing 2 data filtering 2 data quality 2 dataset construction 2 dataset creation 2 dexterous manipulation 2 diffusion transformers 2 direct preference optimization 2 discrete diffusion 2 document retrieval 2 document understanding 2 dynamic adaptation 2 dynamic environments 2 dynamic routing 2 end-to-end training 2 error analysis 2 error detection 2 exploration bottleneck 2 flow matching 2 grpo algorithm 2 gui automation 2 gui grounding 2 hierarchical retrieval 2 hierarchical rl 2 high-resolution generation 2 human mesh recovery 2 image personalization 2 information retrieval 2 iterative optimization 2 iterative refinement 2 kv cache 2 language agents 2 lifelong learning 2 llm integration 2 llm-as-a-judge 2 long context 2 long sequence processing 2 long-context modeling 2 long-context understanding 2 medical diagnosis 2 medical imaging 2 memory architecture 2 mobile automation 2 model calibration 2 model interpretability 2 model quantization 2 monocular depth estimation 2 monte carlo tree search 2 multi-modal evaluation 2 multi-modal learning 2 multi-task learning 2 multi-view consistency 2 multilingual evaluation 2 multimodal conditioning 2 multimodal control 2 multimodal diffusion transformer 2 multimodal models 2 mutual information 2 parallel sampling 2 parallel training 2 parameter-efficient fine-tuning 2 partial observability 2 pomdp framework 2 position encoding 2 privacy protection 2 process supervision 2 progressive learning 2 prompt injection defense 2 quantization 2 question answering 2 reasoning 2 reasoning enhancement 2 reasoning evaluation 2 reasoning optimization 2 reinforcement learning fine-tuning 2 reward design 2 reward model 2 rlvr 2 robustness evaluation 2 safety guardrails 2 self-improvement 2 self-rewarding 2 semantic alignment 2 semi-supervised learning 2 sim-to-real 2 sim2real transfer 2 speech synthesis 2 supervised fine-tuning 2 synthetic dataset 2 test-time adaptation 2 test-time scaling 2 text-to-3d 2 text-to-image diffusion 2 theorem proving 2 token compression 2 training stability 2 tree search 2 ui automation 2 unified framework 2 video editing 2 video quality assessment 2 video synthesis 2 vision transformer 2 vision-language fusion 2 vision-language pre-training 2 visual language models 2 voice generation 2 web navigation 2 zero-shot generalization 2 zero-shot performance 2 3d content generation 1 3d ct analysis 1 3d face modeling 1 3d generation evaluation 1 3d grounding 1 3d hair modeling 1 3d human generation 1 3d human reconstruction 1 3d medical data 1 3d mesh generation 1 3d modeling 1 3d object completion 1 3d occupancy grounding 1 3d perception 1 3d point clouds 1 3d point tracking 1 3d pose estimation 1 3d rendering 1 3d rope 1 3d scene generation 1 3d scene representation 1 3d scene synthesis 1 3d vision 1 4d reconstruction 1 4d scenes 1 4k generation 1 6d video representation 1 academic benchmarking 1 academic conferences 1 academic retrieval 1 academic survey generation 1 accessibility tree pruning 1 accountability 1 acoustic-semantic gap 1 action decoding 1 action generation 1 action relabeling 1 actionable augmentation 1 activation differences 1 activation engineering 1 activation function 1 activation steering 1 active learning 1 adaptive computation 1 adaptive conditioning 1 adaptive exploration 1 adaptive inference 1 adaptive normalization 1 adaptive re-masking 1 adaptive routing 1 adaptive strategy 1 adaptive temperature sampling 1 adaptive thinking 1 adaptive training 1 adaptive training strategy 1 advantage estimation 1 adventure games 1 adversarial attack 1 adversarial evaluation 1 adversarial instructions 1 adversarial testing 1 adversarial training 1 advertising video understanding 1 affective computing 1 agent alignment 1 agent frameworks 1 agent optimization 1 agent safety 1 agent systems 1 agent training 1 agent trajectories 1 agentic continual pre-training 1 agentic enhancement 1 agentic rl 1 agentic search 1 agentic systems 1 ai companionship 1 ai control protocols 1 ai security 1 ai-human collaboration 1 algorithmic alignment 1 alignment 1 alignment evaluation 1 alignment safety 1 alignment-free training 1 ambiguity detection 1 anatomy-aware modeling 1 anime hairstyles 1 anomaly generation 1 answer consistency 1 antisemitism detection 1 api reliability 1 approximate nearest neighbor search 1 arena evaluation 1 arithmetic optimization 1 artistic generation 1 asynchronous architecture 1 asynchronous execution 1 asynchronous rl 1 attack strategies 1 attention analysis 1 attention compression 1 attention heads 1 attention projection 1 attention refinement 1 attribute preservation 1 attribute transfer 1 attribution generation 1 audio benchmark 1 audio compression 1 audio generation 1 audio-language models 1 audio-visual alignment 1 audio-visual dataset 1 audio-visual generation 1 audio-visual reasoning 1 audio-visual synchronization 1 audio-visual-text integration 1 authorship attribution 1 automated annotation 1 automated benchmarking 1 automated data synthesis 1 automated design 1 automated machine learning 1 automated model design 1 automated optimization 1 automated research 1 automated scientific discovery 1 automated strategy development 1 automated testing 1 automated theorem proving 1 automated thinking 1 automated tool generation 1 automated training 1 automatic evaluation 1 automatic rewards 1 autonomous data collection 1 autonomous data generation 1 autonomous systems 1 autoregressive decoder 1 autoregressive diffusion 1 autoregressive prediction 1 autoscaling 1 auxiliary training objective 1 avatar animation 1 background replacement 1 battle prediction 1 bavarian dialect 1 bayesian entropy 1 behavior cloning 1 behavioral benchmarking 1 behavioral fingerprinting 1 behavioral intervention 1 belief conflict index 1 benchmark challenge 1 benchmark creation 1 benchmark paraphrasing 1 benchmark survey 1 benchmarking methodology 1 bfs algorithm 1 bias detection 1 bias mitigation 1 binary classification 1 bio-inspired model 1 biological agents 1 biomedical ner 1 black-box evaluation 1 blind low-vision accessibility 1 block expansion 1 bootstrap sampling 1 brain-inspired architecture 1 brain-inspired memory 1 budget optimization 1 budget-aware inference 1 cad generation 1 camera pose estimation 1 camera trajectory 1 capability evaluation 1 caption generation 1 carbon footprint 1 cascaded framework 1 cascaded generation 1 case-based reasoning 1 catastrophic forgetting 1 causal mask 1 cc-hard benchmark 1 chain-of-human-preference 1 chain-of-thought prompting 1 chart understanding 1 checkpoint reuse 1 chemical reasoning 1 chinese nlp 1 citation verification 1 classifier-free guidance 1 clinical tasks 1 clustering algorithms 1 co-evolution 1 cochlear tokens 1 code as action 1 code assistance 1 code attribution 1 code completion 1 code embeddings 1 code execution verification 1 code generation security 1 code llm 1 code reasoning 1 code retrieval 1 code stylometry 1 code translation 1 code understanding 1 coevolution 1 cognitive architecture 1 cognitive bias 1 cognitive decline 1 cognitive framework 1 cognitive inertia 1 cognitive reasoning 1 cognitive simulation 1 cognitive understanding 1 collaborative training 1 color fidelity 1 community-driven annotation 1 community-driven platform 1 competitive programming 1 complementary control 1 complex logic tasks 1 complex scenes 1 complexity scaling 1 complexity theory 1 compositional design 1 compositional reasoning 1 compositional visual reasoning 1 computational complexity 1 compute scaling 1 concept generation 1 concept removal 1 conditional steering 1 conditional transformer 1 confidence filtering 1 confidence gap 1 consistency models 1 constraint generation 1 containerized testing framework 1 content richness 1 content-style decoupling 1 context collapse 1 context engineering 1 context evolution 1 context length 1 context protocols 1 context summarization 1 context-aware fusion 1 context-aware generation 1 context-aware systems 1 context-free grammar 1 contextual bandits 1 contextual policy optimization 1 continual learning 1 continuous reward functions 1 continuous-time dynamics 1 contrastive identity loss 1 control points 1 controllable reasoning 1 controlled generation 1 controlled text generation 1 controlnet 1 conversational ai 1 conversational framework 1 copy-paste artifacts 1 coset sampling 1 cost efficiency 1 cost optimization 1 cost-of-pass metric 1 cost-performance tradeoff 1 counterintuitive reasoning 1 critic training 1 critic-free rl 1 cross-attention 1 cross-attention heatmaps 1 cross-cultural evaluation 1 cross-domain generalization 1 cross-domain training 1 cross-encoder 1 cross-lingual evaluation 1 cross-lingual generalization 1 cross-modal alignment 1 cross-modal fusion 1 cross-modal integration 1 cross-modal representation 1 cross-model generalization 1 cross-platform generalization 1 cross-resolution 1 cross-view reconstruction 1 crt construction 1 curriculum reinforcement learning 1 cve vulnerability assessment 1 cybersecurity agents 1 dark humor detection 1 data contamination 1 data distribution 1 data flywheel 1 data governance 1 data mixing optimization 1 data provenance 1 data recovery 1 data selection 1 dataset benchmark 1 dataset cleaning 1 dataset quality 1 de-identification 1 debiasing framework 1 decentralization 1 deception benchmark 1 decision tokens 1 decoding acceleration 1 decontextualization 1 decoupled rl 1 deep-research agents 1 deformation modeling 1 degeneration-aware training 1 dehallucination mechanism 1 delimiter brittleness 1 demonstration conditioning 1 dense geometry estimation 1 depth estimation 1 developer tools 1 dialect robustness 1 dialogue state tracking 1 differentiable rendering 1 differential privacy 1 difficulty scaling 1 diffusion acceleration 1 diffusion feature extraction 1 digital agents 1 digital signal processing 1 directed acyclic graph 1 disaggregated architecture 1 discrete operational modes 1 discrete token space 1 discrete tokenization 1 discriminative verification 1 dishonest behavior 1 distributed computing 1 distribution matching 1 diversity optimization 1 diversity preservation 1 docker containers 1 document conversion 1 document memory 1 document parsing 1 document quality assessment 1 domain specialization 1 domain transfer 1 domain-specific llm 1 downstream performance 1 dpo 1 dpo optimization 1 dpo training 1 draw semantics 1 drug discovery 1 drug toxicity prediction 1 dual learning 1 dual-system framework 1 dual-use knowledge 1 dynamic 3d content 1 dynamic advantage function 1 dynamic assembly 1 dynamic benchmark 1 dynamic component optimization 1 dynamic computation allocation 1 dynamic computation efficiency 1 dynamic data selection 1 dynamic dataset 1 dynamic defense 1 dynamic difficulty adjustment 1 dynamic evaluation 1 dynamic fine-tuning 1 dynamic knowledge graphs 1 dynamic memory 1 dynamic part segmentation 1 dynamic pruning 1 dynamic reasoning 1 dynamic reinforcement learning 1 dynamic resource allocation 1 dynamic reward 1 dynamic reward mechanisms 1 dynamic sampling 1 dynamic scenes 1 dynamic source selection 1 dynamic time warping 1 dynamic tokenizer 1 dynamic training 1 dynamic validation 1 dynamic verification 1 dynamic visual reasoning 1 early decoding 1 early exiting 1 early termination 1 echo training 1 edge deployment 1 efficiency 1 efficiency benchmarks 1 efficiency-performance tradeoff 1 efficient architecture 1 efficient fine-tuning 1 efficient transformers 1 embedded deployment 1 embedding refinement 1 embodied reasoning 1 emergent behavior 1 emergent misalignment 1 emotional reasoning 1 end-to-end framework 1 enterprise systems 1 entity embedding 1 entity masking 1 entity tracking 1 entropy balancing 1 entropy collapse 1 entropy modulation 1 entropy optimization 1 entropy regularization 1 environment interaction 1 environment tuning 1 error classification 1 error correction 1 evaluation benchmarks 1 evaluation metrics 1 evaluation paradigm 1 evidence bounds 1 evidence-based reasoning 1 executable environment 1 execution evaluation 1 execution safety 1 experience generation 1 experience knowledge base 1 expert behavior 1 expert iteration 1 expert stitching 1 expert validation 1 expert-amateur model 1 exploration 1 exploration behavior 1 exploration efficiency 1 exploration enhancement 1 exploration-exploitation tradeoff 1 external memory 1 external memory management 1 external modules 1 f-divergence 1 face animation 1 factor-model optimization 1 factual consistency 1 failure modes 1 fairness evaluation 1 fairness in ai 1 false-premise detection 1 fashion ai 1 feature aggregation 1 feature disentanglement 1 feature engineering 1 feature extraction 1 feature fusion 1 feature suppression 1 feature upsampling 1 feed-forward model 1 few-shot learning 1 few-step generation 1 few-step inference 1 film editing 1 financial auditing 1 financial decision making 1 financial domain 1 financial reasoning 1 fine-grained detection 1 fine-grained learning 1 fine-grained localization 1 fine-tuning analysis 1 fisher information matrix 1 fixed corpus evaluation 1 flow distillation 1 flow environment 1 flow models 1 flow-matching 1 formal verification 1 foundation model 1 frequency-aware pruning 1 frontend development 1 gan 1 garment fidelity 1 generalization 1 generalization analysis 1 generation chain-of-thought 1 generative architecture 1 generative classification 1 generative critique 1 generative evaluation 1 generative models 1 generator-validator gap 1 genetic algorithms 1 geo-visual agents 1 geographic information systems 1 geometric constraints 1 geometric deep learning 1 geometric reasoning 1 geometric supervision 1 geometry reasoning 1 german nlp 1 global reasoning 1 goal tracking 1 gradient bias correction 1 gradient preservation 1 gradient redistribution 1 gradient stabilization 1 gradient variance reduction 1 gradient-based evaluation 1 graph transformers 1 greedy objectives 1 group influence metric 1 grpo 1 gui foundation model 1 gui interaction 1 gumbel distribution 1 hair generation 1 hallucination reduction 1 hardware acceleration 1 hardware optimization 1 harm rejection 1 headline generation 1 hessian approximation 1 heterogeneous computing 1 heuristic learning 1 hierarchical assessment 1 hierarchical constraint satisfaction 1 hierarchical construction 1 hierarchical image generation 1 hierarchical indexing 1 hierarchical prompting 1 hierarchical reinforcement learning 1 hierarchical tree 1 high-frequency trading 1 high-resolution image understanding 1 high-resolution training 1 high-resolution video 1 high-risk scenarios 1 hiring assessments 1 human evaluation 1 human interaction animation 1 human motion generation 1 human preference assessment 1 human reasoning styles 1 human-ai interaction 1 human-centered interaction 1 human-centric video 1 human-computer interaction 1 human-robot interaction 1 human-to-robot learning 1 hybrid agents 1 hybrid ensemble 1 hybrid environment 1 hybrid framework 1 identity-consistent generation 1 image inpainting 1 image quality assessment 1 image refinement 1 image synthesis 1 image-text retrieval 1 imitation learning 1 implicit actor-critic 1 implicit world modeling 1 in-tool learning 1 inconsistency resolution 1 incremental compilation 1 industrial inspection 1 industrial migration 1 inference efficiency 1 inference paradigm 1 inference-time optimization 1 information fidelity 1 information gain 1 information gathering 1 inpainting guidance 1 input reformulation 1 input scrutiny 1 instruction augmentation 1 instruction routing 1 instruction-guided image editing 1 instruction-guided video editing 1 interactive agents 1 interactive ai 1 interactive design 1 interactive environment 1 interactive recommendation 1 interactive rewards 1 interactive tasks 1 interactive video generation 1 intermediate representations 1 intermediate tasks 1 internal activations 1 internal mechanisms 1 interpretable ai 1 interpretation quality assessment 1 interruption robustness 1 intuitive physics 1 inverse projection 1 ip-adapter 1 iterative reasoning 1 iupac naming 1 jailbreak defense 1 jailbreak detection 1 javascript code 1 junk text exposure 1 kinematic segmentation 1 knn correlation 1 knowledge alignment 1 knowledge extraction 1 knowledge graph extraction 1 knowledge graphs 1 knowledge reasoning 1 knowledge refinement 1 knowledge representation 1 knowledge retrieval 1 knowledge synthesis 1 knowledge verification 1 knowledge-driven agents 1 knowledge-intensive tasks 1 kv cache compression 1 kv cache optimization 1 kv cache pruning 1 language comprehension 1 language model 1 language model development 1 language model hallucination 1 language model safety 1 language variation 1 large-scale generation 1 last-token pooling 1 last-token reward 1 latency optimization 1 latent diffusion 1 latent space 1 latent space reasoning 1 layer selective optimization 1 layer skipping 1 layout analysis 1 layout optimization 1 layout understanding 1 layout-as-thought 1 length generalization 1 length-aware training 1 lightweight transformer 1 likelihood preference 1 linear attention 1 linear complexity 1 linguistic bias 1 lip synchronization 1 listnet loss 1 listwise ranking 1 llm agents 1 llm as judges 1 llm benchmarking 1 llm constraints 1 llm for science 1 llm optimization 1 llm routing 1 llm serving 1 llm training data 1 llm-driven agents 1 llm-driven extraction 1 load balancing 1 local deepfakes 1 local deployment 1 local search optimization 1 long context modeling 1 long context processing 1 long sequence modeling 1 long video synthesis 1 long video understanding 1 long-chain reasoning 1 long-context 1 long-context evaluation 1 long-context inference 1 long-context processing 1 long-form qa 1 long-horizon execution 1 long-horizon navigation 1 long-tailed learning 1 long-term planning 1 louvain detection 1 low-bit inference 1 low-bit precision 1 low-bitrate compression 1 low-probability regularization 1 low-resource language 1 machine translation 1 machine unlearning 1 macro-level reasoning 1 malicious content filtering 1 masked autoencoder 1 material field 1 math reasoning 1 mathematical problem solving 1 matrix factorization 1 measurement theory 1 medical ai 1 medical image classification 1 medical image retrieval 1 medical image segmentation 1 medical llm 1 medical retrieval 1 memory enhancement 1 memory management 1 memory mechanisms 1 memory safety 1 memory systems 1 memory-augmented agents 1 mesh generation 1 meta-cognition 1 meta-imitation learning 1 meta-reinforcement learning 1 metacognitive reasoning 1 metaphysical shift 1 misinformation resistance 1 mitre att&ck 1 mixture-of-rewards 1 ml research automation 1 mllm 1 mllm evaluation 1 mllms 1 mobile manipulation 1 modality trust 1 model abstraction 1 model clustering 1 model compression 1 model deployment 1 model diagnostics 1 model editing 1 model fingerprinting 1 model optimization 1 model pruning 1 model representation 1 model robustness 1 model safety 1 model stitching 1 model training 1 modular framework 1 module communities 1 moe alternative 1 molecular design 1 molecular representation 1 monocular 3d reconstruction 1 monocular vision 1 multi-agent architecture 1 multi-agent evaluation 1 multi-agent games 1 multi-agent reinforcement learning 1 multi-agent rl 1 multi-agent search 1 multi-agent system 1 multi-answer generation 1 multi-armed bandit 1 multi-asset synthesis 1 multi-component planning 1 multi-dimensional analysis 1 multi-document reasoning 1 multi-file reasoning 1 multi-hop qa 1 multi-human interaction 1 multi-identity datasets 1 multi-knowledge retrieval 1 multi-modal agents 1 multi-modal llm 1 multi-modal perception 1 multi-modal reasoning 1 multi-objective optimization 1 multi-omics integration 1 multi-path reasoning 1 multi-person synthesis 1 multi-policy fusion 1 multi-scale token aggregation 1 multi-scale upsampling 1 multi-speaker 1 multi-stage reasoning 1 multi-stage tasks 1 multi-step refinement 1 multi-supervision learning 1 multi-task evaluation 1 multi-task training 1 multi-teacher learning 1 multi-token prediction 1 multi-tool optimization 1 multi-turn agents 1 multi-turn reinforcement learning 1 multi-turn tool use 1 multi-view fusion 1 multidimensional modeling 1 multidisciplinary benchmark 1 multilingual benchmark 1 multilingual commonsense reasoning 1 multilingual corpus 1 multilingual dataset 1 multilingual language model 1 multilingual llm 1 multilingual nlp 1 multilingual ocr 1 multilingual retrieval 1 multilingual safety 1 multimodal alignment 1 multimodal conditional generation 1 multimodal data 1 multimodal data fusion 1 multimodal dataset 1 multimodal datasets 1 multimodal encoding 1 multimodal feedback 1 multimodal generation 1 multimodal instruction 1 multimodal interaction 1 multimodal learning 1 multimodal llm evaluation 1 multimodal pretraining 1 multimodal retrieval-augmented generation 1 multimodal training 1 multimodal training curriculum 1 multimodal video framework 1 music restoration 1 music theory 1 named entity retrieval 1 narrative analysis 1 narrative audio generation 1 narrative structure 1 native 3d generation 1 native-resolution vision 1 natural language control 1 network analysis 1 neural audio modeling 1 neural compression 1 neural dynamics 1 neural rendering 1 neural simulation 1 neuroscience-inspired ai 1 newton-raphson 1 nlp evaluation 1 no-pair training 1 noise correlation mitigation 1 noise filtering 1 noise inversion 1 noise robustness 1 noise scheduling 1 non-autoregressive generation 1 nonsense understanding 1 numerical methods 1 numerical reasoning 1 object detection 1 object hallucination 1 object recall 1 object removal 1 observation-action gap 1 occlusion handling 1 ocr-free extraction 1 ode sampling 1 off-policy fine-tuning 1 offline rl 1 online adaptation 1 online fine-tuning 1 online tracking 1 ood detection 1 open framework 1 open licensing 1 open-domain qa 1 open-domain tasks 1 optimizer design 1 out-of-distribution 1 over-refusal reduction 1 panoramic image generation 1 paper retrieval 1 paper search agent 1 paper-to-agent 1 paralinguistic tasks 1 parallel computing 1 parallel decoding 1 parallel distillation 1 parallel reasoning 1 parameter expansion 1 parameter merging 1 parameter-efficient rl 1 parametric model 1 parametric modeling 1 parametric models 1 part decomposition 1 partial forgery 1 pass@k 1 pass@k optimization 1 pass@k training 1 passage ranking 1 patch discriminators 1 pathology diagnosis 1 patient simulator 1 peer evaluation 1 perceptual optimization 1 performance enhancement 1 performance gap 1 performance metrics 1 performance saturation 1 perlin noise 1 personality analysis 1 personalization bias 1 personalized safety alignment 1 perspective taking 1 persuasive dialogues 1 perturbation prediction 1 phase-aware rewards 1 physical constraints 1 physical property prediction 1 physical realism 1 physical simulation 1 pii redaction 1 planner-executor framework 1 planning evaluation 1 player identification 1 point cloud autoencoder 1 point cloud completion 1 point cloud generation 1 policy composition 1 policy gradient 1 policy gradient optimization 1 policy gradients 1 policy learning 1 portrait animation 1 pose estimation 1 pose priors 1 pose retargeting 1 post-training 1 post-training quantization 1 potential space 1 ppo 1 ppo optimization 1 pre-execution safety 1 pre-training 1 pre-training efficiency 1 pre-training optimization 1 preference aggregation 1 preference decomposition 1 preference elicitation 1 preference hijacking 1 preference modeling 1 primitive-mesh decomposition 1 privacy awareness 1 privacy benchmark 1 privacy-preserving ml 1 proactive ai 1 probabilistic framework 1 probability density distillation 1 procedural content generation 1 procedural knowledge 1 process annotation 1 process reasoning reward 1 process reward model 1 process reward modeling 1 process reward models 1 production systems 1 programming assistants 1 programming language 1 programming reasoning 1 progress reward 1 progressive mesh generation 1 prompt engineering 1 prompt injection 1 prompt optimization 1 prompt sensitivity 1 pseudo-annotation dataset 1 pseudo-thompson sampling 1 psychological computing 1 psychological safety 1 pure reinforcement learning 1 python interpreter 1 qa data synthesis 1 qft 1 quadrotor control 1 quality-diversity tradeoff 1 quantitative analysis 1 quantitative finance 1 quantum algorithms 1 query analysis 1 query rewriting 1 query routing 1 quiz-based evaluation 1 radiology reports 1 rag 1 rag framework 1 rag limitations 1 rag systems 1 rare disease 1 rating systems 1 real-time inference 1 real-time systems 1 real-world data 1 reasoning analysis 1 reasoning benchmarks 1 reasoning capabilities 1 reasoning chains 1 reasoning dataset 1 reasoning decomposition 1 reasoning diversity 1 reasoning framework 1 reasoning leakage 1 reasoning modes 1 reasoning queries 1 reasoning skill taxonomy 1 reasoning sparks 1 reasoning styles 1 reasoning synthesis 1 reasoning-intensive domains 1 reasoning-intensive tasks 1 rectified flow matching 1 recursive transformers 1 redundancy reduction 1 reference image injection 1 reference-based inpainting 1 reflection behavior 1 reflective agents 1 refusal tuning 1 reinforce algorithm 1 reinforce++ 1 reinforcement fine-tuning 1 reinforcement learning alignment 1 reinforcement learning optimization 1 reinforcement mid-training 1 relighting 1 replay learning 1 report quality 1 repository-level evaluation 1 repository-level pretraining 1 representation alignment 1 representation autoencoder 1 representation bias 1 representation learning 1 representation steering 1 research synthesis 1 research systems 1 residual learning 1 residual policy 1 resilient ai 1 resolution-agnostic 1 resource optimization 1 responsible ai 1 retrieval benchmarks 1 retrieval models 1 retrieval optimization 1 retrieval systems 1 retrieval-augmented simulation 1 reverse-engineered reasoning 1 reward fine-tuning 1 reward function 1 reward hacking 1 reward optimization 1 reward shaping 1 reward variance 1 reward-guided decoding 1 rhetorical analysis 1 risk assessment 1 risk detection 1 risk management 1 rl fine-tuning 1 rl for reasoning 1 rl optimization 1 rlhf 1 robot interaction 1 robotic reasoning 1 robotics interaction 1 robustness 1 rope fine-tuning 1 rpa scripts 1 rubric-scaffolded rl 1 safe agi 1 safe alignment 1 safety benchmark 1 safety constraints 1 safety filtering 1 safety preservation 1 sample filtering 1 sample selection 1 sampling efficiency 1 scale-aligned reference 1 scene composition 1 scienceboard benchmark 1 scientific analysis 1 scientific discovery 1 scientific publishing 1 scientific reasoning 1 score distillation 1 search agents 1 search intelligence 1 security vulnerability 1 selective refusal 1 self-alignment 1 self-conditioning 1 self-consistency 1 self-consistency voting 1 self-correction 1 self-critique 1 self-evolving data 1 self-evolving reasoning 1 self-improving language models 1 self-improving models 1 self-play 1 self-reflection 1 self-reflective ai 1 self-rewarding models 1 self-supervised pre-training 1 self-verification 1 semantic advantage 1 semantic classifier 1 semantic coherence 1 semantic consistency 1 semantic correctness 1 semantic disambiguation 1 semantic divergence 1 semantic equivalence 1 semantic evaluation 1 semantic features 1 semantic fragility 1 semantic matching 1 semantic preserving rewrite 1 semantic prompts 1 semantic reasoning 1 semantic segmentation 1 semantic stability 1 semantic tree 1 semi-structured data 1 sentiment analysis 1 sequence analysis 1 sequence-to-sequence 1 service framework 1 shap analysis 1 side effect handling 1 simultaneous speech translation 1 single-cell biology 1 single-cell omics 1 single-image 3d 1 single-stage training 1 skeleton-shape disentanglement 1 sketch-based generation 1 skill targeting 1 skill-based finetuning 1 slam 1 slowfast encoding 1 small language models 1 smartphone agents 1 social behavior understanding 1 social bias 1 social deduction games 1 social intelligence 1 social media analysis 1 social reasoning 1 soft q-learning 1 software development automation 1 software engineering 1 software issue localization 1 solution filtering 1 source-free learning 1 span-level annotation 1 sparse action enhancement 1 sparse autoencoders 1 sparse computation 1 sparse models 1 sparse networks 1 sparse reward 1 sparse view reconstruction 1 sparsification 1 spatial encoding 1 spatial generalization 1 spatial intelligence 1 spatial reasoning 1 spatial transformer 1 spatial understanding 1 spatio-temporal understanding 1 specialized llm 1 speculative decoding 1 speech emotion recognition 1 speech generation 1 speech llm 1 speech representation learning 1 speech tokenization 1 speech-language model 1 stage-wise optimization 1 stance dynamics 1 state preparation 1 state transition 1 static sparsity 1 statistical learning theory 1 step accuracy 1 step verification 1 stepwise reasoning 1 stochastic sub-networks 1 strategic planning 1 streaming detection 1 streaming efficiency 1 structure preservation 1 structured curriculum 1 structured data 1 structured data analysis 1 structured evaluation 1 structured knowledge system 1 structured output 1 structured task design 1 structured text evaluation 1 structured visual representation 1 style formalization 1 style professionalism 1 style transfer 1 stylized rendering 1 subject-driven generation 1 subtask verification 1 summary generation failure 1 super-resolution 1 supervised learning 1 supervised rl 1 surface defect detection 1 surface normal estimation 1 surprisal metric 1 survey 1 sustainability 1 sustainable ai 1 sycophancy detection 1 symbolic music reasoning 1 synthetic media 1 system 1 system 2 1 system 2 thinking 1 system documentation 1 system error 1 system identification 1 system-algorithm co-design 1 systematic generalization 1 table question answering 1 talking head generation 1 tamper-resistance 1 task alignment 1 task generation 1 task performance 1 task planning 1 template optimization 1 temporal consistency 1 temporal decoupling 1 temporal dynamics 1 temporal grounding 1 temporal synchronization 1 tensor parallelism 1 test-time optimization 1 test-time search 1 text augmentation 1 text embeddings 1 text encoder optimization 1 text preprocessing 1 text-aware modeling 1 text-based person re-id 1 text-controlled audio 1 text-driven generation 1 text-guided image editing 1 text-image fusion 1 text-to-image evaluation 1 text-to-motion 1 text-to-video editing 1 textual action representation 1 theoretical analysis 1 thinking trajectories 1 thought process generation 1 thought skipping 1 thought templates 1 threat intelligence 1 three-stage training 1 thurstone model 1 token budget 1 token order prediction 1 token pruning 1 token selection 1 token sparsification 1 token-level diffusion 1 tokenization drift 1 tokenizer optimization 1 tool abstraction 1 tool calling 1 tool coordination 1 tool learning 1 tool misuse 1 tool usage strategy 1 tool-augmented agents 1 tool-augmented verification 1 tool-integrated reasoning 1 tool-integrated training 1 tool-use agents 1 topology optimization 1 training architecture 1 training data generation 1 training data synthesis 1 training efficiency 1 training paradigm 1 training-free 1 training-free optimization 1 trajectory alignment 1 trajectory attention 1 trajectory filtering 1 trajectory generation 1 trajectory optimization 1 transformer 1 transformer acceleration 1 transformer compression 1 transformer modeling 1 transparency framework 1 triplane diffusion 1 trust evaluation 1 trusted monitoring 1 two-stage training 1 ui agents 1 ui simulation 1 uncertain reasoning 1 uncertainty estimation 1 uncertainty handling 1 uncertainty quantification 1 uncertainty-aware ranking 1 understanding-generation alignment 1 unified generation model 1 unified training framework 1 unsolved questions 1 unsupervised adaptation 1 unsupervised domain adaptation 1 unsupervised segmentation 1 user embedding 1 user interface 1 user memory 1 user-centric evaluation 1 validation mechanism 1 value estimation 1 variance-aware sampling 1 variational problem synthesis 1 vector graphics 1 velocity deviation 1 velocity matching 1 vertex splitting 1 video alignment 1 video captioning 1 video dataset 1 video diffusion 1 video editing benchmark 1 video generation benchmark 1 video inpainting 1 video manipulation detection 1 video models 1 video object segmentation 1 video processing 1 video reasoning 1 video relighting 1 video segmentation 1 video state prediction 1 video super-resolution 1 video temporal reasoning 1 video-based world models 1 video-language model 1 video-language models 1 video-to-4d 1 view consistency 1 viewpoint planning 1 virtual cell modeling 1 virtual social agents 1 virtual world generation 1 vision foundation models 1 vision transformers 1 vision-grounded rl 1 vision-language alignment 1 visual action prompts 1 visual agent 1 visual autoregressive modeling 1 visual chain of thought 1 visual data synthesis 1 visual game generation 1 visual generation 1 visual hallucination reduction 1 visual hallucination suppression 1 visual navigation 1 visual noise suppression 1 visual programmability 1 visual quality enhancement 1 visual quality metrics 1 visual resolution routing 1 visual spatial reasoning 1 visual structure 1 visual token masking 1 visual tokenization 1 visual tool use 1 visual uncertainty 1 visual-language-action 1 visual-symbolic reasoning 1 visual-text retrieval 1 visualization system 1 vlm automation 1 vlm feedback 1 voice style adaptation 1 von neumann architecture 1 voxel perception 1 web design to code 1 web interaction 1 world model 1 world modeling 1 xbrl 1 zero-shot adaptation 1 zero-shot reranking 1 zero-shot tools 1 zero-shot tts 1
📄

2509.01656

🤖 系统
09-09 20:56