What Do You See in Common? Learning Hierarchical Prototypes over Tree-of-Life to Discover Evolutionary Traits
Comments ICLR 2025
期刊&会议
ICLR
International Conference on Learning Representations · 会议 · Machine Learning
Operator Deep Smoothing for Implied Volatility
Journal ref The Thirteenth International Conference on Learning Representations, ICLR 2025, Singapore, April 24-28, 2025
On the Completeness of Invariant Geometric Deep Learning Models
机构 * Institute for Artificial Intelligence, Peking University（人工智能研究院，北京大学） ； School of Intelligence Science and Technology, Peking University（智能科学与技术学院，北京大学）
Comments The Thirteenth International Conference on Learning Representations
MogaNet: Multi-order Gated Aggregation Network
机构 * Zhejiang University, College of Computer Science and Technology（浙江大学计算机科学与技术学院） ； AI Lab, Research Center for Industries of the Future（未来产业研究中心人工智能实验室）
Comments ICLR 2024. Preprint V4 (35 pages, fixed typos). Code and models refer to https://github.com/Westlake-AI/MogaNet
SSLAM: Enhancing Self-Supervised Models with Audio Mixtures for Polyphonic Soundscapes
机构 * Surrey Institute for People-Centred AI（萨里人本人工智能研究所） ； University of Surrey（萨里大学） ； Centre for Vision, Speech and Signal Processing (CVSSP)（视觉、语音与信号处理中心）
Comments Accepted at ICLR 2025. Code and pre-trained models are available at \url{https://github.com/ta012/SSLAM}
ViSAGe: Video-to-Spatial Audio Generation
机构 * Seoul National University（首尔国立大学）
Comments ICLR 2025. Project page: https://jaeyeonkim99.github.io/visage/
AegisLLM: Scaling Agentic Systems for Self-Reflective Defense in LLM Security
机构 * University of Maryland College Park（马里兰大学学院公园分校） ； Lawrence Livermore National Laboratory（劳伦斯利弗莫尔国家实验室） ； Capital One
Comments ICLR 2025 Workshop BuildingTrust
A Causal Lens for Learning Long-term Fair Policies
机构 * Department of Electrical Engineering and Computer Science（电气工程与计算机科学系） ； University of Arkansas（阿肯色大学）
Comments This is an extension to the paper which was accepted to the 13th International Conference on Learning Representations
Geometrically Aligned Transfer Encoder for Inductive Transfer in Regression Tasks
机构 * LG AI Research（LG人工智能研究）
Comments 12+11 pages, 6+1 figures, 0+7 tables
Journal ref ICLR 2024 Main Conference
Task Adaptation from Skills: Information Geometry, Disentanglement, and New Objectives for Unsupervised Reinforcement Learning
机构 * Eindhoven University of Technology（埃因霍温理工大学） ； University of Maryland, College Park（马里兰大学学院市分校） ； Ruhr University Bochum（波鸿鲁尔大学） ； Tencent Robotics X（腾讯机器人科技） ； University of Liverpool（利物浦大学）
Comments Spotlight paper at ICLR 2024. This version includes acknowledgments omitted from the ICLR version and indicates the corresponding authors primarily responsible for the work
Journal ref International Conference on Learning Representations (ICLR), 2024, Spotlight paper
TexTailor: Customized Text-aligned Texturing via Effective Resampling
机构 * KAIST Daejeon, South Korea（韩国庆北国立大学）
Comments Submitted to ICLR 2025
Measuring Semantic Information Production in Generative Diffusion Models
机构 * Department of Artificial Intelligence（人工智能系） ； Radboud University（拉德堡德大学） ； Department of Information Technology（信息技术系） ； Ghent University（根特大学） ； Jheronimus Academy of Data Science（杰赫罗尼穆斯数据科学学院） ； Tilburg University（蒂尔堡大学） ； Donders Institute for Brain, Cognition and Behaviour（大脑、认知与行为研究所）
Comments 4 pages, 3 figures, an appendix with derivations and implementation details, accepted at ICLR DeLTa 2025
Amulet: ReAlignment During Test Time for Personalized Preference Adaptation of LLMs
Comments Accepted by ICLR 2025, Project page: https://zowiezhang.github.io/projects/Amulet
MGMapNet: Multi-Granularity Representation Learning for End-to-End Vectorized HD Map Construction
机构 * Tongji University（同济大学） ； Baidu Inc.（百度公司）
Journal ref https://www.iclr.cc/virtual/2025/poster/30420
FinTSBridge: A New Evaluation Suite for Real-world Financial Prediction with Advanced Time Series Models
机构 * Tsinghua University（清华大学） ； Peng Cheng Laboratory（鹏城实验室）
Comments ICLR 2025 Workshop Advances in Financial AI
Towards Foundational Models for Dynamical System Reconstruction: Hierarchical Meta-Learning via Mixture of Experts
机构 * School of Computer Science, University of Bristol, UK.（计算机科学学院，布里斯托大学） ； School of Physics, University of Bristol, UK.（物理学院，布里斯托大学） ； School of Engineering Mathematics and Technology, University of Bristol, UK.（工程数学与技术学院，布里斯托大学）
Comments 22 pages, 11 figures, 7 tables. Accepted as a SCOPE workshop paper at ICLR 2025
Journal ref https://openreview.net/forum?id=NteuHm0UXw (2025)
Understanding Long Videos with Multimodal Language Models
Comments 17 pages (main paper), 7 pages appendix. ICLR 2025 conference paper
Generating Likely Counterfactuals Using Sum-Product Networks
机构 * Department of Computer Science Faculty of Electrical Engineering, Czech Technical University（计算机科学系电子工程学院，捷克技术大学）
Comments 32 pages total
Journal ref The Thirteenth International Conference on Learning Representations (ICLR 2025)
Dynamic Negative Guidance of Diffusion Models
Comments Paper accepted at ICLR 2025 (poster). Our implementation is available at https://github.com/FelixKoulischer/Dynamic-Negative-Guidance.git
Effective Data Augmentation With Diffusion Models
机构 * Carnegie Mellon University（卡内基梅隆大学） ； MPG Ranch（MPG牧场） ； University Of Chicago（芝加哥大学） ； Laboratory Schools（实验室学校）
Comments Update to ICLR 2024 manuscript (https://openreview.net/forum?id=ZWzUA9zeAg), add leafy spurge citations
Offline RL with Smooth OOD Generalization in Convex Hull and its Neighborhood
机构 * School of Mathematical Sciences, Beihang University（北京航空航天大学数学科学学院） ； School of Physics, Beihang University（北京航空航天大学物理学院） ； School of Artificial Intelligence, Beihang University（北京航空航天大学人工智能学院） ； Key Laboratory of Mathematics, Informatics and Behavioral Semantics, MoE, Beihang University（北京航空航天大学数学、信息学与行为语义重点实验室） ； Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing, Beihang University（北京未来区块链与隐私计算先进创新中心） ； Zhongguancun Laboratory（中关村实验室） ； Hangzhou International Innovation Institute of Beihang University（杭州北京航空航天大学国际创新研究院）
Comments ICLR 2025
FairMT-Bench: Benchmarking Fairness for Multi-turn Dialogue in Conversational LLMs
机构 * Zhejiang University（浙江大学） ； Zhejiang Key Laboratory of Medical Imaging Artificial Intelligence（浙江省医学影像人工智能重点实验室）
Comments ICLR 2025 spotlight
FlickerFusion: Intra-trajectory Domain Generalizing Multi-Agent RL
机构 * Yonsei University（延世大学） ； KAIST AI（韩国科学技术院人工智能研究所）
Comments ICLR 2025
Stabilizing Contrastive RL: Techniques for Robotic Goal Reaching from Offline Data
Comments ICLR 2024 Spotlight (< 5%). Website (https://chongyi-zheng.github.io/stable_contrastive_rl) and code (https://github.com/chongyi-zheng/stable_contrastive_rl)
AstroCompress: A benchmark dataset for multi-purpose compression of astronomical data
机构 * UC Irvine（uci大学） ； UC Berkeley（伯克利大学） ； LBNL（劳伦斯伯克利国家实验室）
Comments ICLR 2025 conference paper. See reviews at https://openreview.net/forum?id=kQCHCkNk7s
Regretful Decisions under Label Noise
Comments The Thirteenth International Conference on Learning Representations (ICLR 2025)
QP-SNN: Quantized and Pruned Spiking Neural Networks
机构 * University of Electronic Science and Technology of China（电子科学与技术大学） ； Northumbria University（诺斯姆比亚大学） ； Liaoning Technical University（辽宁技术大学）
Comments 26 pages, 17 figures, Published as a conference paper at ICLR 2025
Hummingbird: High Fidelity Image Generation via Multimodal Context Alignment
机构 * Microsoft（微软公司） ； Stony Brook University（史蒂文尼森布鲁克大学）
Comments Accepted to ICLR 2025. Project page with code release: https://roar-ai.github.io/hummingbird
SMI-Editor: Edit-based SMILES Language Model with Fragment-level Supervision
机构 * School of Computer Science, Peking University（北京大学计算机科学学院） ； National Key Laboratory for Multimedia Information Processing, Peking University（北京大学多媒体信息处理国家重点实验室） ； Peking University-Anker Embodied AI Lab, Peking University（北京大学安克具身AI实验室） ； International Digital Economy Academy (IDEA), Shenzhen, China（国际数字经济学院（IDEA），深圳，中国） ； College of Computer Science, Sichuan University（四川大学计算机学院） ； Computer Science and Engineering Department, University of Washington（华盛顿大学计算机科学与工程系）
Comments ICLR 2025
Dynamic-SUPERB Phase-2: A Collaboratively Expanding Benchmark for Measuring the Capabilities of Spoken Language Models with 180 Tasks
Comments ICLR 2025
