ICML 2024 Papers — Page 27
International Conference on Machine Learning · 2610 papers
WISER: Weak Supervision and Supervised Representation Learning to Improve Drug Response Prediction in Cancer
Kumar Shubham (Indian Institute of Science), Vaibhav Rajan (National University of Singapore)
Representation LearningDrug DiscoveryGenerative Adversarial NetworkContrastive LearningBiomedical Data
🎯 What it does: Developed the WISER framework, which combines weak supervision with supervised domain-invariant representation learning to improve drug response prediction in cancer patients.
WorkArena: How Capable are Web Agents at Solving Common Knowledge Work Tasks?
Alexandre Drouin (ServiceNow Research), Alexandre Lacoste (Mila - Quebec AI Research Institute)
TransformerLarge Language ModelTextMultimodalityBenchmarkChain-of-Thought
🎯 What it does: The WorkArena benchmark is proposed to evaluate common knowledge work tasks driven by large language models in the enterprise software ServiceNow, and a BrowserGym environment is constructed to support multimodal observations and various actions.
Wukong: Towards a Scaling Law for Large-Scale Recommendation
Buyun Zhang (Meta AI), Wenlin Chen (Meta AI)
Recommendation SystemTabular
🎯 What it does: The Wukong network architecture is proposed, utilizing stacked Factorization Machines (FM) to capture arbitrary-order feature interactions, and achieving a sustainably improved recommendation model through 'dense expansion';
X-Oscar: A Progressive Framework for High-quality Text-guided 3D Animatable Avatar Generation
Yiwei Ma (Xiamen University), Rongrong Ji (Xiamen University)
GenerationData SynthesisOptimizationDiffusion modelScore-based ModelTextMesh
🎯 What it does: Proposes the X-Oscar framework, which generates high-quality, animatable 3D avatars using text prompts.
xT: Nested Tokenization for Larger Context in Large Images
Ritwik Gupta (University of California), Karttikeya Mangalam (University of California)
ClassificationObject DetectionSegmentationTransformerImage
🎯 What it does: By using nested tokenization and a two-stage pipeline, the existing visual models that only process small images can integrate large-scale context of the entire image while maintaining local details, achieving classification, detection, and segmentation of large images.
Zero-Shot ECG Classification with Multimodal Learning and Test-time Clinical Knowledge Enhancement
Che Liu (Imperial College London), Rossella Arcucci (Imperial College London)
ClassificationRepresentation LearningConvolutional Neural NetworkLarge Language ModelContrastive LearningMultimodalityBiomedical DataElectrocardiogram
🎯 What it does: We propose a multi-modal learning-based ECG representation learning framework MERL, which can achieve zero-shot ECG classification by utilizing cross-modal alignment between ECG signals and corresponding reports.
Zero-Shot Reinforcement Learning via Function Encoders
Tyler Ingebrand (University of Texas at Austin), ufuk topcu
Reinforcement Learning
🎯 What it does: This paper proposes a Function Encoder to represent reward or transition functions as a weighted combination of several learned nonlinear basis functions, resulting in a unique vector representation, which is then used as contextual input for any reinforcement learning algorithm to achieve zero-shot transfer.
Zero-Shot Unsupervised and Text-Based Audio Editing Using DDPM Inversion
Hila Manor (Technion Israel Institute of Technology), Tomer Michaeli (Technion Israel Institute of Technology)
GenerationData SynthesisDiffusion modelAudio
🎯 What it does: Two zero-shot audio editing methods are proposed, namely text-based ZETA and unsupervised ZEUS, utilizing pre-trained diffusion models for audio editing.
Zero-Sum Positional Differential Games as a Framework for Robust Reinforcement Learning: Deep Q-Learning Approach
Anton Plaksin (Yandex), Vitaly Kalev (IMM UB RAS)
Reinforcement Learning
🎯 What it does: This paper proposes to view the robust reinforcement learning problem as a differential game, utilizing the Isaacs condition to prove that the shared Q-function can simultaneously satisfy the minimax and maximin Bellman equations. Based on this, two deep Q-network methods, IDQN and DIDQN, are designed to address continuous high-dimensional robust RL tasks.
Zeroth-Order Methods for Constrained Nonconvex Nonsmooth Stochastic Optimization
Zhuanghua Liu (National University of Singapore), Bryan Kian Hsiang Low (National University of Singapore)
Recommendation SystemOptimizationTabular
🎯 What it does: This paper addresses the solution of non-convex, non-smooth stochastic optimization problems on closed convex sets, proposing two classes of zero-order (gradient-free) stochastic projection gradient descent and Frank-Wolfe algorithms, and provides non-asymptotic convergence analysis.