Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks544 from public data
- CoolStreaming/DONet: A Data-Driven Overlay Network for Efficient Live Media Streaming635
This paper presents DONet, a Data-driven Overlay Network for live media streaming, and presents an efficient memberand partnership management algorithm, together with an intelligent scheduling algorithm that achieves real-time and continuous distribution of streaming contents.
- OpenOOD: Benchmarking Generalized Out-of-Distribution Detection416
A unified, well-structured codebase called OpenOOD is built, which implements over 30 methods developed in relevant fields and provides a comprehensive benchmark under the recently proposed generalized OOD detection framework.
- Google USM: Scaling Automatic Speech Recognition Beyond 100 Languages412
The Universal Speech Model is introduced, a single large model that performs automatic speech recognition (ASR) across 100+ languages by pre-training the encoder of the model on a large unlabeled multilingual dataset, and fine-tuning on a smaller labeled dataset.
- 399
- Efficient task-specific data valuation for nearest neighbor algorithms317
An Monte Carlo approximation algorithm is proposed, which is up to three orders of magnitude faster than the baseline approximation algorithm and can accelerate the value calculation process even further.
- MIMIC-IT: Multi-Modal In-Context Instruction Tuning316
MultI-Modal In-Context Instruction Tuning (MIMIC-IT), a dataset comprising 2.8 million multimodal instruction-response pairs, with 2.2 million unique instructions derived from images and videos, is presented and a large VLM named Otter is trained.
- 275
- Federated Learning with Label Distribution Skew via Logits Calibration271
This paper investigates the label distribution skew in FL and proposes FedLC, which calibrates the logits before softmax cross-entropy according to the probability of occurrence of each class, and applies a fine-grained calibrated cross-entropy loss to local update by adding a pairwise label margin.
- Video-MMMU: Evaluating Knowledge Acquisition from Multi-Discipline Professional Videos241
Video-MMMU is introduced, a multi-modal, multi-disciplinary benchmark designed to assess LMMs' ability to acquire and utilize knowledge from videos, and proposes a proposed knowledge gain metric, {\Delta}knowledge, that quantifies improvement in performance after video viewing.
- Evaluating ChatGPT's Information Extraction Capabilities: An Assessment of Performance, Explainability, Calibration, and Faithfulness221
ChatGPT's performance in Standard-IE setting is poor, but it surprisingly exhibits excellent performance in the OpenIE setting, as evidenced by human evaluation, and indicates that ChatGPT provides high-quality and trustworthy explanations for its decisions.
- HALC: Object Hallucination Reduction via Adaptive Focal-Contrast Decoding220
HALC, a novel decoding algorithm designed to mitigate OH in LVLMs, is introduced, which leverages distinct fine-grained optimal visual information in vision-language tasks and operates on both local and global contexts simultaneously.
- Cross-Modal Progressive Comprehension for Referring Segmentation202
This paper proposes a cross-modal progressive comprehension (CMPC) scheme to effectively mimic human behaviors and implements it as a CMPC-I (Image) module and aCMPC-V (Video) module to improve referring image and video segmentation models.
- UI-TARS-2 Technical Report: Advancing GUI Agent with Multi-Turn Reinforcement Learning201
UI-TARS-2 is presented, a native GUI-centered agent model that addresses challenges through a systematic training methodology: a data flywheel for scalable data generation, a stabilized multi-turn RL framework, a hybrid GUI environment that integrates file systems and terminals, and a unified sandbox platform for large-scale rollouts.
- PixMix: Dreamlike Pictures Comprehensively Improve Safety Measures200
A new data augmentation strategy utilizing the natural structural complexity of pictures such as fractals is designed, which outperforms numerous baselines, is near Pareto-optimal, and roundly improves safety measures.
- Adversary Guided Asymmetric Hashing for Cross-Modal Retrieval167
This paper proposes a novel deep cross-modal hashing method, called Adversary Guided Asymmetric Hashing (AGAH), which employs an adversarial learning guided multi-label attention module to enhance the feature learning part which can learn discriminative feature representations and keep the cross- modal invariability.
- 165
- LINKTELLER: Recovering Private Edges from Graph Neural Networks via Influence Analysis144
This paper proposes a privacy attack LINKTELLER via influence analysis to infer the private edge information held by Alice via designing adversarial queries for Bob, and empirically shows that LINK TELLER is able to recover a significant amount of private edges in different settings, both including inductive and transductive datasets, under different graph densities, significantly outperforming existing baselines.
- On the Tool Manipulation Capability of Open-source Large Language Models139
It is demonstrated that classical methods in LLM literature can adapt as model alignment with programmatic data generation, system prompts and in-context demonstration retrievers to enhance open-source LLMs for tool manipulation, and it is shown that such enhancement typically requires about one developer day to curate data for each tool, rendering a recipe with practical amount of human supervision.
- MixGen: A New Multi-Modal Data Augmentation136
MixGen is presented: a joint data augmentation for vision-language representation learning to further improve data efficiency and generates new image-text pairs with semantic relationships preserved by interpolating images and concatenating text.
- LSTM-Based Analysis of Industrial IoT Equipment135
This paper aims to develop a method of analyzing equipment working condition based on the sensed data and building a prediction model for working status forecasting and designing a deep neural network model to predict equipment running data and improving the prediction accuracy by systematic feature engineering and optimal hyperparameter searching.
- Step-Audio: Unified Understanding and Generation in Intelligent Speech Interaction129
Step-Audio is introduced, the first production-ready open-source solution that achieves state-of-the-art performance in human evaluations, especially in terms of instruction following and on open-source benchmarks like LLaMA Question.
- DiffScene: Diffusion-Based Safety-Critical Scenario Generation for Autonomous Vehicles128
A unified approach to generate high-quality safety-critical scenarios which are both realistic and safety-critical for efficient AV evaluation, and it is demonstrated that training AV algorithms with scenarios generated by DiffScene leads to significantly higher performance in terms of the safety-critical metrics compared to baselines.
- 128
- A comparison of methods for sketch-based 3D shape retrieval127
A more comprehensive comparison of fifteen best sketch-based 3D shape retrieval methods is performed by completing the evaluation of each method on both benchmarks.
- Feature extraction using constrained maximum variance mapping126
Experimental results validate that the proposed CMVM approach is superior to other feature extraction methods, such as linear discriminant analysis (LDA), locality preserving projection (LPP), unsupervised discriminant projection (UDP) and maximum variance projection (MVP).
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-14. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.