ICML 2024

2,660 papers

Top Research Directions

Research directions most represented by ICML 2024 papers.

Overcoming the Iterative Inference Latency of Diffusion Models 24
Safe Reinforcement Learning via Constrained Markov Decision Processes 21
The Non-Linear Function Approximation Gap in Reinforcement Learning 21
Sample Efficient Reinforcement Learning 21
Clean Accuracy Degradation in Machine Learning Defenses 21
Overcoming the Homophily Assumption in Graph Neural Networks 18
The Mechanics of Backdoor Attacks in Neural Networks 17
The Weisfeiler Lehman Expressivity Bottleneck 16
The Function Approximation Gap in Reinforcement Learning Theory 16
Mitigating Gradient Variance in Policy Optimization 14
Matrix-Free Curvature Estimation for Deep Learning 14
Optimization and Generalization with Non-Smooth Activations 14
Accelerating Sparse Neural Networks on Dense Hardware 14
Scaling Gaussian Processes Beyond Cubic Complexity 14
Bypassing the Computational Bottleneck of Adversarial Training 14

Top Authors

AuthorPapers
Bo Li12
James Y. Zou12
Jianye HAO12
Pin-Yu Chen11
Tongliang Liu11
James Zou10
Junchi Yan10
Sergey Levine10
Stan Z. Li10
Volkan Cevher10
Zhangyang Wang10
Dacheng Tao9
Furong Huang9
Rongrong Ji9
Taiji Suzuki9
Andrew Gordon Wilson8
Bo Han8
David Woodruff8
Dawn Song8
De-Chuan Zhan8

Impactful Papers

Papers from ICML 2024 that shaped research clusters.

Foundational

Influential

Papers

Title
Position: Application-Driven Innovation in Machine Learning
Position: A Roadmap to Pluralistic Alignment
Position: A Safe Harbor for AI Evaluation and Red Teaming
Position: Automatic Environment Shaping is the Next Frontier in RL
Position: Bayesian Deep Learning is Needed in the Age of Large-Scale AI
Position: Benchmarking is Limited in Reinforcement Learning Research
Position: Beyond Personhood: Agency, Accountability, and the Limits of Anthropomorphic Ethical Analysis
Position: Building Guardrails for Large Language Models Requires Systematic Design
Position: C∗-Algebraic Machine Learning - Moving in a New Direction
Position: Categorical Deep Learning is an Algebraic Theory of All Architectures
Position: Compositional Generative Modeling: A Single Model is Not All You Need
Position: Considerations for Differentially Private Learning with Large-Scale Public Pretraining
Position: Cracking the Code of Cascading Disparity Towards Marginalized Communities
Position: Data Authenticity, Consent, & Provenance for AI are all broken: what will it take to fix them?
Position: Data-driven Discovery with Large Generative Models
Position: Do Not Explain Vision Models Without Context
Position: Do pretrained Transformers Learn In-Context by Gradient Descent?
Position: Embracing Negative Results in Machine Learning
Position: Enforced Amnesia as a Way to Mitigate the Potential Risk of Silent Suffering in the Conscious AI
Position: Evolving AI Collectives Enhance Human Diversity and Enable Self-Regulation
Position: Explain to Question not to Justify
Position: Exploring the Robustness of Pipeline-Parallelism-Based Decentralized Training
Position: Foundation Agents as the Paradigm Shift for Decision Making
Position: Fundamental Limitations of LLM Censorship Necessitate New Approaches
Position: Future Directions in the Theory of Graph Machine Learning
Position: Graph Foundation Models Are Already Here
Position: Insights from Survey Methodology can Improve Training Data
Position: Intent-aligned AI Systems Must Optimize for Agency Preservation
Position: Is machine learning good or bad for the natural sciences?
Position: Key Claims in LLM Research Have a Long Tail of Footnotes
Position: Levels of AGI for Operationalizing Progress on the Path to AGI
Position: Leverage Foundational Models for Black-Box Optimization
Position: Lifetime tuning is incompatible with continual reinforcement learning
Position: LLMs Can’t Plan, But Can Help Planning in LLM-Modulo Frameworks
Position: Machine Learning-powered Assessments of the EU Digital Services Act Aid Quantify Policy Impacts on Online Harms
Position: Measure Dataset Diversity, Don't Just Claim It
Position: Mission Critical – Satellite Data is a Distinct Modality in Machine Learning
Position: Near to Mid-term Risks and Opportunities of Open-Source Generative AI
Position: On the Possibilities of AI-Generated Text Detection
Position: On the Societal Impact of Open Foundation Models
Position: Open-Endedness is Essential for Artificial Superhuman Intelligence
Position: Opportunities Exist for Machine Learning in Magnetic Fusion Energy
Position: Optimization in SciML Should Employ the Function Space Geometry
Position Paper: Bayesian Deep Learning in the Age of Large-Scale AI
Position: Quo Vadis, Unsupervised Time Series Anomaly Detection?
Position: Reinforcement Learning in Dynamic Treatment Regimes Needs Critical Reexamination
Position: Relational Deep Learning - Graph Representation Learning on Relational Databases
Position: Rethinking Post-Hoc Search-Based Neural Approaches for Solving Large-Scale Traveling Salesman Problems
Position: Scaling Simulation is Neither Necessary Nor Sufficient for In-the-Wild Robot Manipulation
Position: Scarce Resource Allocations That Rely On Machine Learning Should Be Randomized

« prev Page 37/54 next »