Crossref 2025

1,192 papers
Crossref 2025 featured 993 papers, with a significant emphasis on the safety, reliability, and architectural constraints of large language models (LLMs). Key research themes included active context management for long-horizon agents, modeling language within finite working memory, and the representational alignment of artificial and biological neural networks. Safety-related studies were prominent, focusing on adversarial intent extraction, latent space monitoring, and the development of low-latency guardrails to ensure generative security. Additionally, research addressed factual faithfulness in abstractive summarization and the grounding of models for clinical applications. Technical contributions also explored privacy-preserving cloud-device collaboration, positional debiasing for context window extension, and iterative model-assisted data annotation. Notable work included the development of the OxBioZ wearable bioimpedance monitoring system and decentralized application upgrade systems using blockchain. Prolific authors contributing to this edition included Maarten de Rijke, Julian McAuley, and Dawei Yin. The year's output reflected a broader trend toward enhancing model robustness and practical utility in specialized domains such as healthcare, hierarchical clustering, and generative model aesthetic evaluation.

Top Research Directions

Research directions most represented by Crossref 2025 papers.

Overcoming the Limits of Multi Step Language Model Reasoning 5
Self Supervised Regularization for Sparse Recommender Systems 5
Dynamic Preference Modeling in Interactive Systems 5
Factual Inaccuracy and Hallucination in Large Language Models 5
Reducing the Memory and Latency of Language Model Inference 5
Individualized Instruction for Heterogeneous Learners 4
Intermediate Reasoning Chains for Deliberative Inference 4
Clean Accuracy Degradation in Machine Learning Defenses 4
Generative Recommendation with Large Language Models 3
Advanced Representations and Denoising for Robust Recommendation 3
Training Models via Progressive Task Complexity 3
Breaking the Accuracy Diversity Dilemma in Recommendations 3
Grounding Language Models with External Knowledge Retrieval 3
Cross-Modal Attention Fusion for Heterogeneous Data Integration 3
Overcoming the Static Graph Assumption 3

Top Authors

AuthorPapers
Maarten de Rijke26
Julian McAuley13
Shuaiqiang Wang11
Dawei Yin10
Shazia Sadiq10
Flora D. Salim9
Juho Kim9
Simon S. Woo9
Xiangyu Zhao9
Enhong Chen8
Hao Xue8
Qingyao Ai8
Quan Z. Sheng8
Fabian J Theis7
Haochen You7
Imran Razzak7
KC Santosh7
Rodrigo Alves7
Ruiming Tang7
Zhaochun Ren7

Impactful Papers

Papers from Crossref 2025 that shaped research clusters.

Foundational

Influential

Papers

Title
Textualized and Feature-Based Models for Compound Multimodal Emotion Recognition in the Wild
Textual-Visual Logic Challenge: Understanding and Reasoning in Text-to-Image Generation
Textual-Visual Logic Challenge: Understanding and Reasoning in Text-to-Image Generation
TFcomb identifies transcription factor combinations for cellular reprogramming based on single-cell multiomics data
TGIF2 is a major regulator of neural stem cell fate and neurogenic priming
The 1 st NIP@IR Workshop on New Interaction Paradigms for Information Retrieval in the Era of Generative AI
The 3rd Workshop on UAVs in Multimedia: Capturing the World from a New Perspective
The 4th Workshop on AI Agent for Information Retrieval: Generating and Ranking
The Birth of Vision Language
The Data Behind the Model: Gaps and Opportunities for Foundation Models in Brain Imaging
The Data Behind the Model: Gaps and Opportunities for Foundation Models in Brain Imaging
The Future is Sparse: Embedding Compression for Scalable Retrieval in Recommender Systems
The HCI GenAI CO2ST Calculator: Calculating and Offsetting the Carbon Footprint of Generative AI Use in Human-Computer Interaction Research
The Hitchhiker's Guide to Autonomous Research: A Survey of Scientific Agents
The Mediating Role of Structural Connectivity in Genetic Effects on Functional Brain Networks
The Omniroute maze: a novel rodent navigation apparatus that integrates dynamically configurable routes, sensory cues and automated reward delivery
The perceptual primacy of feeling: Affectless machine vision models explain a majority of variance in human visually-evoked affect
The price of flexibility in electricity markets
The Procedural Content Generation Benchmark: An Open-source Testbed for Generative Challenges in Games
The Second Search Futures Workshop at ECIR’25
The Second Search Futures Workshop at ECIR’25
The Second Workshop on Generative AI for E-commerce
The Topological Architecture of Brain Identity
The Unlikely Hero: Nonidealities in Analog Photonic Neural Networks as Built-in Adversarial Defenders
The World Wide recipe: A community-centred framework for fine-grained data collection and regional bias operationalisation
Think Together and Work Better: Combining Humans' and LLMs' Think-Aloud Outcomes for Effective Text Evaluation
Through the Lens of Privacy: Exploring Privacy Protection in Vision-Language Model Interactions on Smart Glasses
Tight Results for Online Convex Paging
Time Matters: Enhancing Sequential Recommendations with Time-Guided Graph Neural ODEs
Time Series Analysis in the Web: Recent Advances and Future Trends
TimeSoccer: An End-to-End Multimodal Large Language Model for Soccer Commentary Generation
TIM-Rec: Explicit Sparse Feedback on Multi-Item Upselling Recommendations in an Industrial Dataset of Telco Calls
Toward a Training-Free Plug-and-Play Refinement Framework for Infrared and Visible Image Registration and Fusion
Toward Holistic Evaluation of Recommender Systems Powered by Generative Models
Towards a Bass Diffusion Forecasting Approach for the UK's Smart Meter Roll-Out
Towards a New Paradigm of Visual Signal Compression
Towards Comprehensive Legal Document Analysis: A Multi-Round RAG Approach
Towards Diagnostic Quality Flat-Panel Detector CT Imaging Using Diffusion Models
Towards Efficient Few-shot Graph Neural Architecture Search via Partitioning Gradient Contribution
Towards End-to-End Optimization of LLM-based Applications with Ayo
Towards Harmless Multimodal Assistants with Blind Preference Optimization
Towards Machine-Generated Code for the Resolution of User Intentions
Towards Modality Generalization: A Benchmark and Prospective Analysis
Towards Non-Euclidean Foundation Models: Advancing AI Beyond Euclidean Frameworks
Towards Open Domain-Specific Recognition using Quad-Channel Self-Attention Reciprocal Point Learning and Autoencoder
Towards Reliable Latent Knowledge Estimation in LLMs: Zero-Prompt Many-Shot Based Factual Knowledge Extraction
Towards Robust and Controllable Text-to-Motion via Masked Autoregressive Diffusion
Towards Safe Synthetic Image Generation On the Web: A Multimodal Robust NSFW Defense and Million Scale Dataset
Towards Secure AI in Education: A Case Study on Automatic Short Answer Grading
Towards Secure AI in Education: A Case Study on Automatic Short Answer Grading

« prev Page 22/24 next »