ablation studies
Experiments that systematically remove or alter components of a model to assess their impact on performance, helping to understand which parts are most crucial.
- A Clean Slate for Offline Reinforcement Learning
- AceReason-Nemotron: Advancing Math and Code Reasoning through Reinforcement Learning
- Active Target Discovery under Uninformative Priors: The Power of Permanent and Transient Memory
- AltLoRA: Towards Better Gradient Approximation in Low-Rank Adaptation with Alternating Projections
- Automated Model Discovery via Multi-modal & Multi-step Pipeline
- Bifrost-1: Bridging Multimodal LLMs and Diffusion Models with Patch-level CLIP Latents
- C-LoRA: Contextual Low-Rank Adaptation for Uncertainty Estimation in Large Language Models
- CAT: Circular-Convolutional Attention for Sub-Quadratic Transformers
- DSAS: A Universal Plug-and-Play Framework for Attention Optimization in Multi-Document Question Answering
- Decoupling Contrastive Decoding: Robust Hallucination Mitigation in Multimodal Large Language Models
- Dual-Res Tandem Mamba-3D: Bilateral Breast Lesion Detection and Classification on Non-contrast Chest CT
- Dual-Space Semantic Synergy Distillation for Continual Learning of Unlabeled Streams
- Embodied Cognition Augmented End2End Autonomous Driving
- Enhancing Multilingual LLM Pretraining with Model-Based Data Selection
- Exploiting the Asymmetric Uncertainty Structure of Pre-trained VLMs on the Unit Hypersphere
- Frame Context Packing and Drift Prevention in Next-Frame-Prediction Video Diffusion Models
- Holistic Order Prediction in Natural Scenes
- How Far Are We from Optimal Reasoning Efficiency?
- IPSI: Enhancing Structural Inference with Automatically Learned Structural Priors
- Improving Retrieval-Augmented Generation through Multi-Agent Reinforcement Learning
- InfoChartQA: A Benchmark for Multimodal Question Answering on Infographic Charts
- LeVo: High-Quality Song Generation with Multi-Preference Alignment
- Learning Chern Numbers of Multiband Topological Insulators with Gauge Equivariant Neural Networks
- Learning to Rank for In-Context Example Retrieval
- MEIcoder: Decoding Visual Stimuli from Neural Activity by Leveraging Most Exciting Inputs
- MLE-STAR: Machine Learning Engineering Agent via Search and Targeted Refinement
- MODEM: A Morton-Order Degradation Estimation Mechanism for Adverse Weather Image Recovery
- Mellow: a small audio language model for reasoning
- Memory-Integrated Reconfigurable Adapters: A Unified Framework for Settings with Multiple Tasks
- MetaMind: Modeling Human Social Thoughts with Metacognitive Multi-Agent Systems
- Model Merging in Pre-training of Large Language Models
- MokA: Multimodal Low-Rank Adaptation for MLLMs
- MokA: Multimodal Low-Rank Adaptation for MLLMs
- NUTS: Eddy-Robust Reconstruction of Surface Ocean Nutrients via Two-Scale Modeling
- OCN: Effectively Utilizing Higher-Order Common Neighbors for Better Link Prediction
- Online Feedback Efficient Active Target Discovery in Partially Observable Environments
- Open-Reasoner-Zero: An Open Source Approach to Scaling Up Reinforcement Learning on the Base Model
- PT-MoE: An Efficient Finetuning Framework for Integrating Mixture-of-Experts into Prompt Tuning
- Personalized Decision Modeling: Utility Optimization or Textualized-Symbolic Reasoning
- ReMA: Learning to Meta-Think for LLMs with Multi-agent Reinforcement Learning
- Real-DRL: Teach and Learn in Reality
- Reasoning Is Not a Race: When Stopping Early Beats Going Deeper
- Robust Label Proportions Learning
- Scaling Computer-Use Grounding via User Interface Decomposition and Synthesis
- Self-Training with Dynamic Weighting for Robust Gradual Domain Adaptation
- Towards Realistic Earth-Observation Constellation Scheduling: Benchmark and Methodology
- UniGen: Enhanced Training & Test-Time Strategies for Unified Multimodal Understanding and Generation
- Unifying Attention Heads and Task Vectors via Hidden State Geometry in In-Context Learning
- Vulnerable Data-Aware Adversarial Training