Conference on Neural Information Processing Systems · 1376 papers
Text Alignment Is An Efficient Unified Model for Massive NLP Tasks
Yuheng Zha (University of California San Diego), Zhiting Hu (University of California San Diego)
CodeClassificationData SynthesisTransformerLarge Language ModelSupervised Fine-TuningText
🎯 What it does: A text alignment model (ALIGN) is proposed, which unifies various NLP tasks into a task that measures the degree of information alignment between two pieces of text.
Text Promptable Surgical Instrument Segmentation with Vision-Language Models
Zijian Zhou (King's College London), Miaojing Shi (Tongji University)
CodeObject DetectionSegmentationTransformerPrompt EngineeringVision Language ModelImageMultimodality
🎯 What it does: A text prompt-based surgical instrument segmentation method is proposed, transforming the segmentation task into an image + text input, utilizing a vision-language model for fine recognition of different instruments.
Michael Hassid (Hebrew University of Jerusalem), Yossi Adi (Hebrew University of Jerusalem)
CodeRecognitionGenerationTransformerLarge Language ModelBenchmarkAudio
🎯 What it does: This paper proposes a warm-init method called TWIST for Speech Language Model (SpeechLM) based on a pre-trained text language model, and trains the largest SpeechLM (7B/13B) model on a large-scale speech dataset, while releasing two versions of the speech-based StoryCloze benchmark.
TFLEX: Temporal Feature-Logic Embedding Framework for Complex Reasoning over Temporal Knowledge Graph
Xueyuan Lin (Beijing University of Posts and Telecommunications), Mingzhi Sun (Beijing University of Posts and Telecommunications)
CodeGraph Neural NetworkGraphTime Series
🎯 What it does: This paper proposes the TFLEX framework, which can perform multi-hop logical reasoning on temporal knowledge graphs (TKG), supporting entity queries and temporal queries, covering all first-order logical operations as well as temporal operations such as After, Before, and Between.
The Behavior and Convergence of Local Bayesian Optimization
Kaiwen Wu (University of Pennsylvania), Jacob R. Gardner (University of Pennsylvania)
CodeOptimization
🎯 What it does: This paper conducts a systematic study of Local Bayesian Optimization (LBO), providing empirical experimental validation that local optimization can achieve very good local solutions on high-dimensional Gaussian process (GP) sample paths, and also presents a rigorous convergence rate proof for the recently proposed GIBO algorithm in both noise-free and noisy environments.
The Benefits of Being Distributional: Small-Loss Bounds for Reinforcement Learning
Kaiwen Wang (Cornell University), Wen Sun (Cornell University)
CodeReinforcement LearningTabular
🎯 What it does: This paper provides theoretical guarantees for distributed reinforcement learning (DistRL) in the context of small losses, and proposes corresponding online and offline algorithms as well as a distributed contextual bandit (DistCB).
The Clock and the Pizza: Two Stories in Mechanistic Explanation of Neural Networks
Ziqian Zhong (Massachusetts Institute of Technology), Jacob Andreas (Massachusetts Institute of Technology)
CodeTransformerTabular
🎯 What it does: The study investigates how neural networks implement the known Clock algorithm and the new Pizza algorithm when training for modular addition, and explores the phase transitions of the algorithms under different hyperparameters.
The Contextual Lasso: Sparse Linear Models via Deep Neural Networks
Ryan Thompson (University of New South Wales), Robert Kohn (University of New South Wales)
CodeSupervised Fine-TuningTabular
🎯 What it does: This paper proposes a context-sparse linear model and the corresponding Context Lasso estimator, utilizing feedforward neural networks to learn context features and sparse coefficient functions, and implementing a projection layer to impose a constraint on the average ℓ₁ norm, balancing interpretability and expressive power.
The Emergence of Essential Sparsity in Large Pre-trained Models: The Weights that Matter
AJAY KUMAR JAISWAL, Zhangyang Wang (University of Texas at Austin)
CodeTransformerLarge Language ModelVision Language ModelMultimodality
🎯 What it does: This study investigates the phenomenon of 'essential sparsity' in large pre-trained visual and language Transformers, which maintains high performance even after a single round of direct pruning.
The expressive power of pooling in Graph Neural Networks
Filippo Maria Bianchi (UiT Arctic University of Norway), Veronica Lachi (University of Siena)
CodeGraph Neural NetworkGraph
🎯 What it does: This study investigates the expressive power of pooling layers in graph neural networks and provides sufficient conditions for maintaining the expressiveness of MP layers, verifying whether different pooling methods meet these conditions. Additionally, a synthetic dataset EXPWL1 based on the WL test is proposed for empirical testing of the expressiveness of pooling layers.
The Goldilocks of Pragmatic Understanding: Fine-Tuning Strategy Matters for Implicature Resolution by LLMs
Laura Eline Ruis, Edward Grefenstette (University College London)
CodeTransformerLarge Language ModelSupervised Fine-TuningPrompt EngineeringTextChain-of-Thought
🎯 What it does: This study constructs a binary implicature resolution task and systematically evaluates various large language models (baseline, dialogue fine-tuning, benchmark instruction fine-tuning, example instruction fine-tuning) to explore their reasoning abilities regarding conversational implicature.
The Grand Illusion: The Myth of Software Portability and Implications for ML Progress.
Fraser Mince (Cohere for AI Community), Sara Hooker (Cohere for AI)
CodeBenchmark
🎯 What it does: Evaluate the portability and performance differences between TensorFlow, PyTorch, and JAX on GPU and TPU, and release a public test set.
The Graph Pencil Method: Mapping Subgraph Densities to Stochastic Block Models
Lee M. Gunderson (University College London), Peter Orbanz (University College London)
CodeGraph Neural NetworkGraph
🎯 What it does: The Graph Pencil Method is proposed, which accurately maps the densities of finite star and double star subgraphs to the parameters of the Stochastic Block Model (SBM) through root subgraph algebra and matrix pencil techniques, achieving direct inference with no significant computational overhead.
The Memory-Perturbation Equation: Understanding Model's Sensitivity to Data
Peter Nickl (RIKEN Center for AI Project), Mohammad Emtiyaz Khan (RIKEN Center for AI Project)
CodeOptimizationConvolutional Neural NetworkImage
🎯 What it does: Proposes the Memory-Perturbation Equation (MPE), which quickly estimates the model's sensitivity to training data through natural gradient and uses this estimate to predict generalization performance during training.
🎯 What it does: By constructing the Pick-to-Learn (P2L) meta-algorithm, any learning algorithm is transformed into a learning process with compression properties, thereby obtaining compact generalization bounds and improving post-training performance.
Eric J Michaud, Max Tegmark (Massachusetts Institute of Technology)
CodeTransformerLarge Language ModelText
🎯 What it does: A quantization model is proposed to explain the power-law decline of neural network scaling and the sudden emergence of new capabilities.
The Rank-Reduced Kalman Filter: Approximate Dynamical-Low-Rank Filtering In High Dimensions
Jonathan Schmidt (University of Tübingen), Filip Tronarp (Lund University)
CodeTime SeriesStochastic Differential Equation
🎯 What it does: A low-rank Kalman filter (RRKF) is proposed, which achieves approximate Gaussian filtering and posterior estimation in high-dimensional state spaces by maintaining a low-rank approximation of the covariance.
The Rashomon Importance Distribution: Getting RID of Unstable, Single Model-based Variable Importance
Jon Donnelly (Duke University), Edward P Browne
CodeBiomedical Data
🎯 What it does: This study investigates how to quantify variable importance through the calculation of Rashomon Importance Distribution (RID) in the presence of the Rashomon effect and data instability.
The Rise of AI Language Pathologists: Exploring Two-level Prompt Learning for Few-shot Weakly-supervised Whole Slide Image Classification
Linhao Qu (Fudan University), Zhijian Song (Fudan University)
CodeClassificationTransformerLarge Language ModelPrompt EngineeringImageBiomedical Data
🎯 What it does: With a small amount of bag-level annotations, the problem of few-shot weakly supervised classification of whole slide images (WSI) is addressed through dual-layer prompt learning.
The Utility of “Even if” Semifactual Explanation to Optimise Positive Outcomes
Eoin M. Kenny (Massachusetts Institute of Technology), Weipeng Fuzzy Huang
CodeRecommendation SystemAnomaly DetectionOptimizationExplainability and InterpretabilityTabularFinance Related
🎯 What it does: A framework based on 'even if...' counterfactual explanations is proposed and implemented to optimize the benefits of positive decisions made by models (such as loan approvals), rather than just explaining negative outcomes.
Daniel Augusto de Souza (University College London), César Lincoln Mattos
CodeExplainability and InterpretabilityComputational EfficiencyTabular
🎯 What it does: A new high-order Gaussian process model, called Thin and Deep GP (TDGP), is proposed to overcome the limitations of existing deep Gaussian process models, maintaining interpretability and learning low-dimensional embeddings.
🎯 What it does: Proposes the Thinker algorithm, allowing RL agents to autonomously interact with the learned world model and perform planning, integrating model inference and decision-making;
This Looks Like Those: Illuminating Prototypical Concepts Using Multiple Visualizations
Chiyu Ma (Dartmouth), Cynthia Rudin (Duke)
CodeClassificationExplainability and InterpretabilityConvolutional Neural NetworkImage
🎯 What it does: This paper proposes the ProtoConcepts method, which transforms the single image prototype in traditional prototype networks into a multi-image visual concept, using spherical prototypes to capture similar features of multiple training samples, thereby achieving a more intuitive 'this class versus those' explanation.
🎯 What it does: This study investigates the three-way trade-off in multi-objective learning (optimization, generalization, conflict avoidance), proposes the MoDo dual-sampling algorithm, and provides theoretical analysis and experimental validation.
CodeClassificationRetrievalTransformerLarge Language ModelTextRetrieval-Augmented Generation
🎯 What it does: This paper studies how to adaptively retrieve external knowledge in large-scale pre-trained language models (PTLMs), only performing retrieval when the internal knowledge of the model is insufficient, and proposes an instance-level metric called Thrust to evaluate the adequacy of knowledge for a specific instance.
TIES-Merging: Resolving Interference When Merging Models
Prateek Yadav (University of North Carolina at Chapel Hill), Mohit Bansal (University of North Carolina at Chapel Hill)
CodeTransformerSupervised Fine-TuningText
🎯 What it does: A new model merging method called TIES-Merging is proposed, which can merge multiple task-specific fine-tuned models into a multi-task model without additional training, significantly reducing performance loss caused by parameter interference.
Time-uniform confidence bands for the CDF under nonstationarity
Paul Mineiro (Microsoft Research), Steven R Howard
CodeTime Series
🎯 What it does: This paper proposes a time-uniform and value-uniform confidence band for the cumulative distribution function (CDF) of non-stationary, data-dependent univariate sequences, and extends this method to importance-weighted counterfactual distribution estimation.
🎯 What it does: A transferable accelerated molecular dynamics (MD) sampling method named Timewarp is proposed, which generates large-scale structural changes by learning time-coarse-grained dynamics, and then uses Metropolis-Hastings correction within a Markov Chain Monte Carlo (MCMC) framework to sample the thermal equilibrium distribution in an unbiased manner.
TMT-VIS: Taxonomy-aware Multi-dataset Joint Training for Video Instance Segmentation
rongkun Zheng, Hengshuang Zhao (University of Hong Kong)
CodeObject DetectionSegmentationTransformerVideo
🎯 What it does: This paper proposes a model named TMT-VIS, which improves the video instance segmentation task of joint training on multiple datasets by utilizing taxonomy (category hierarchy) information.
🎯 What it does: This paper studies the performance improvement of a model ensemble using only a single pre-trained checkpoint in the context of transfer learning, and proposes a parallel version of Snapshot Ensembling (StarSSE) to better explore model diversity within the pre-trained basin.
Token-Scaled Logit Distillation for Ternary Weight Generative Language Models
Minsoo Kim (Hanyang University), Jungwook Choi (Hanyang University)
CodeGenerationKnowledge DistillationTransformerLarge Language ModelSupervised Fine-TuningText
🎯 What it does: For generative language models (GLMs), a method for quantization-aware training (QAT) under low precision (especially ternary quantization) is proposed, named Token-Scaled Logit Distillation (TSLD).
ToolkenGPT: Augmenting Frozen Language Models with Massive Tools via Tool Embeddings
Shibo Hao (University of California San Diego), Zhiting Hu (University of California San Diego)
CodeGenerationRobotic IntelligenceTransformerLarge Language ModelSupervised Fine-TuningPrompt EngineeringText
🎯 What it does: This paper proposes ToolkenGPT, which allows a frozen LLM to call tools during text generation in the same way as generating ordinary words by learning a token (toolken) embedding for each tool.
🎯 What it does: In response to the weak segmentation of curved structures, this paper proposes a topology-based structural-level uncertainty quantification method;
🎯 What it does: A self-supervised learning framework named TopoSRL has been developed, specifically designed for simplicial complexes, capable of capturing higher-order interactions and preserving topological information in the absence of labeled data.
CodeRecognitionRetrievalTransformerPrompt EngineeringVision Language ModelImageMultimodality
🎯 What it does: The task of 'Re-identify Any Animal in the Wild (ReID-AW)' is proposed, and a general ReID model capable of handling any wild animal category is constructed.
🎯 What it does: A theoretical framework for Generative Data Augmentation (GDA) is proposed, deriving the generalization error upper bound under non-i.i.d. conditions using algorithm stability methods, and applying it to specific analyses of Gaussian Mixture Models and Generative Adversarial Networks (GANs), along with experimental validation.
Towards a fuller understanding of neurons with Clustered Compositional Explanations
Biagio La Rosa (Sapienza University of Rome), Roberto Capobianco (Sony AI)
CodeExplainability and InterpretabilityComputational EfficiencyConvolutional Neural NetworkImage
🎯 What it does: The Clustered Compositional Explanations (CCE) algorithm is proposed, which generates logical formula explanations for neurons by clustering unit activations and using an improved heuristic search (MMESH) on each cluster, providing explanations over a broader range of activations. Based on this, a systematic analysis of the phenomena of 'non-specificity', 'gradual specificity', and 'ambiguity' of neurons under different activation intervals is conducted.
Towards Anytime Classification in Early-Exit Architectures by Enforcing Conditional Monotonicity
Metod Jazbec (University of Amsterdam), Eric Nalisnick (University of Amsterdam)
CodeClassificationMixture of ExpertsImage
🎯 What it does: A post-hoc Product-of-Experts (PoE) transformation is proposed, which allows early exit neural networks to exhibit conditional monotonicity in computation time (i.e., the quality of predictions does not decrease at each step).
Towards Automated Circuit Discovery for Mechanistic Interpretability
Arthur Conmy (Independent), Adrià Garriga-Alonso (FAR AI)
CodeExplainability and InterpretabilityTransformerGraph
🎯 What it does: This paper proposes an automated circuit discovery method called ACDC, which helps researchers identify subgraphs (circuits) in Transformer models that implement specific behaviors.
Towards Better Dynamic Graph Learning: New Architecture and Unified Library
Le Yu (Beihang University), Weifeng Lv (Beihang University)
CodeRepresentation LearningGraph Neural NetworkTransformerGraphTime Series
🎯 What it does: A dynamic graph learning architecture based on Transformer, DyGFormer, and a unified duration graph learning library, DyGLib, are proposed to achieve reproducible, efficient, and scalable dynamic graph representation learning.
Towards Data-Agnostic Pruning At Initialization: What Makes a Good Sparse Mask?
Hoang Pham (FPT Software AI Center), Long Tran-Thanh (University of Warwick)
CodeOptimizationImage
🎯 What it does: The study investigates methods for network pruning during the initialization phase and proposes an evaluation framework based on subnetwork topology.
🎯 What it does: In the context of long-tail imbalance and open-world scenarios, a distribution-independent universal category discovery task that simultaneously handles known and unknown categories is proposed, along with the BaCon framework based on adversarial co-learning.
🎯 What it does: A transfer attack benchmark named TA-Bench is proposed and implemented, covering over 30 existing transfer attack methods and conducting unified evaluations on 25 mainstream victim models.
🎯 What it does: A data selection method called FreeSel is designed, which does not require task-specific models, is unsupervised, and extracts semantic patterns through a pre-trained visual Transformer for distance sampling, achieving efficient data labeling across different visual tasks.
🎯 What it does: A general semi-supervised framework (Aggregating & Decoupling, A&D) is proposed for volumetric medical image segmentation, capable of handling four scenarios: SSL, class-imbalanced SSL, UDA, and SemiDG.
🎯 What it does: The paper proposes a rank-based adversarial weight pruning method (RPG) that enhances model performance by maintaining a high rank of the sparse weight matrix under high sparsity rates.
Towards Hybrid-grained Feature Interaction Selection for Deep Sparse Network
Fuyuan Lyu (McGill University), Xue Liu (McGill University)
CodeRecommendation SystemOptimizationTabular
🎯 What it does: A mixed granularity (field-level and value-level) feature interaction selection method called OptFeature is proposed and implemented in a Deep Sparse Network (DSN).
Towards Label Position Bias in Graph Neural Networks
Haoyu Han (Michigan State University), Jiliang Tang (Michigan State University)
CodeGraph Neural NetworkGraph
🎯 What it does: This paper reveals and mitigates the Label Position Bias in graph neural networks, proposing a metric based on the Label Proximity Score (LPS) and enhancing model performance and fairness through Label Position Unbiased Structure Learning (LPSL).
🎯 What it does: This paper explores how to utilize the visual foundation models CLIP and SAM to achieve label-free 2D/3D scene understanding, and proposes a Cross-Modal Noise Supervision (CNS) framework.
Towards Last-layer Retraining for Group Robustness with Fewer Annotations
Tyler LaBonte (Georgia Institute of Technology), Abhishek Kumar (Google DeepMind)
CodeSupervised Fine-TuningTextBenchmark
🎯 What it does: The study investigates the last layer retraining method in the absence of group labels and proposes the SELF technique with no group labels and only a small number of category labels.
Towards Optimal Caching and Model Selection for Large Model Inference
Banghua Zhu (University of California Berkeley), Jiantao Jiao (University of California Berkeley)
CodeOptimizationLarge Language ModelText
🎯 What it does: A joint optimization framework that combines caching and model multiplexing is proposed to reduce the inference cost and latency of large models.
Towards Personalized Federated Learning via Heterogeneous Model Reassembly
Jiaqi Wang (Pennsylvania State University), Fenglong Ma (Pennsylvania State University)
CodeFederated LearningKnowledge DistillationImage
🎯 What it does: A personalized federated learning framework called pFedHR is proposed, which addresses the collaboration and personalization issues arising from the heterogeneity of model structures across different clients.
🎯 What it does: The RoboSMPLX framework is proposed to achieve full-body pose and shape estimation from monocular images, enhancing robustness and expressiveness.
Towards robust and generalizable representations of extracellular data using contrastive learning
Ankit Vishnubhotla (Columbia University), Cole Lincoln Hurwitz
CodeClassificationRepresentation LearningTransformerContrastive LearningBiomedical Data
🎯 What it does: The CEED framework is proposed, which generates robust features through contrastive learning of extracellular waveforms for spike sorting and cell type classification.
Towards Self-Interpretable Graph-Level Anomaly Detection
Yixin Liu (Monash University), Shirui Pan (Griffith University)
CodeAnomaly DetectionExplainability and InterpretabilityGraph Neural NetworkContrastive LearningGraph
🎯 What it does: This paper studies the interpretable layer anomaly detection model SIGNET, which can provide anomaly scores and subgraph explanations under unsupervised conditions.
Towards Semi-Structured Automatic ICD Coding via Tree-based Contrastive Learning
Chang Lu (Stevens Institute of Technology), Yue Ning (Stevens Institute of Technology)
CodeClassificationContrastive LearningTextBiomedical DataElectronic Health Records
🎯 What it does: This paper proposes an automatic ICD code assignment method based on semi-structured clinical notes. It first automatically extracts and segments the titles and paragraphs of the notes, and then performs contrastive learning and masked training at the paragraph level to reduce note variability and enhance multi-label ICD coding performance.
🎯 What it does: A backdoor removal method based on Feature Shift Tuning (FST) is proposed, specifically targeting backdoor attacks on deep learning models in low contamination scenarios.
Towards Symmetry-Aware Generation of Periodic Materials
Youzhi Luo (Texas A&M University), Shuiwang Ji (Texas A&M University)
CodeGenerationOptimizationGraph Neural NetworkDiffusion modelScore-based ModelAuto EncoderGraphPhysics Related
🎯 What it does: A framework named SyMat for generating symmetric perception periodic materials is proposed, capable of generating chemical compositions, lattice parameters, and atomic coordinates that satisfy physical symmetry constraints.
Meghdad Kurmanji (University of Warwick), Eleni Triantafillou (Google DeepMind)
CodeOptimizationSafty and PrivacyConvolutional Neural NetworkContrastive LearningImage
🎯 What it does: This study investigates the 'forgetting' function of deep learning models and proposes a general forgetting method called SCRUB to address three practical needs: debiasing, eliminating label confusion, and user privacy.
Tracr: Compiled Transformers as a Laboratory for Interpretability
David Lindner (Google DeepMind), Vladimir Mikulik (Google DeepMind)
CodeCompressionExplainability and InterpretabilityTransformerTabular
🎯 What it does: Tracr is proposed and implemented, which is a tool that compiles human-readable RASP programs into Transformer weights, capable of generating Transformer models with known structures for interpretability experiments and evaluations.
Trade-off Between Efficiency and Consistency for Removal-based Explanations
Yifan Zhang (Tsinghua University), Yang Yuan (Tsinghua University)
CodeExplainability and InterpretabilityComputational EfficiencyImageText
🎯 What it does: This paper studies explanation methods based on feature removal and proposes the Incompatibility Triplet Theorem, indicating that interpretability, efficiency, and consistency cannot be satisfied simultaneously; under this framework, a lower error interpreter is designed.
Train Faster, Perform Better: Modular Adaptive Training in Over-Parameterized Models
Yubin Shi (Fudan University), Li Shang (Fudan University)
CodeTransformerImageText
🎯 What it does: This paper studies the modular-level learning dynamics of parameterized models, proposing a modular neural tangent kernel (mNTK) and designing a modular adaptive training (MAT) strategy based on its principal eigenvalue.
Train Once, Get a Family: State-Adaptive Balances for Offline-to-Online Reinforcement Learning
Shenzhi Wang (Tsinghua University), Gao Huang (Tsinghua University)
CodeReinforcement LearningTabularBenchmark
🎯 What it does: This paper proposes the FamO2O framework, which addresses the distribution shift problem in offline to online reinforcement learning by training various conservative/aggressive strategies and adaptively selecting the appropriate strategy based on the state.
Training Chain-of-Thought via Latent-Variable Inference
Du Phan (Google), Rif A. Saurous (Google)
CodeOptimizationComputational EfficiencyTransformerLarge Language ModelPrompt EngineeringTextChain-of-Thought
🎯 What it does: This paper proposes a training framework called TRICE based on MCMC-EM, which automatically generates effective reasoning chains by maximizing the marginal log-likelihood of answers in Chain of Thought (CoT), reducing reliance on manually annotated reasoning chains.
Training neural operators to preserve invariant measures of chaotic attractors
Ruoxi Jiang (University of Chicago), Rebecca Willett (University of Chicago)
CodeContrastive LearningTime SeriesPhysics Related
🎯 What it does: This paper proposes training neural operators through optimal transport or contrastive learning to maintain the invariant measure of chaotic attractors across multiple environments, thereby improving long-term statistical properties.
🎯 What it does: A method for simulating low-resolution image preprocessing of the human retinal periphery, called R-Blur, is proposed, and a DNN is trained based on this to enhance robustness against adversarial attacks and common noise.
Training Private Models That Know What They Don’t Know
Stephan Rabanser (University of Toronto), Nicolas Papernot (Google DeepMind)
CodeClassificationSafty and PrivacyImage
🎯 What it does: This paper studies how to perform Selective Classification (SC) under the constraints of Differential Privacy (DP) and evaluates the performance of different SC methods under various privacy budgets through experiments.
Transfer learning for atomistic simulations using GNNs and kernel mean embeddings
John Isak Texas Falk, Massimiliano Pontil (Istituto Italiano di Tecnologia)
CodeGraph Neural NetworkGraphPhysics Related
🎯 What it does: A transfer learning method based on kernel ridge regression (MEKRR) is proposed to model the potential energy surface of atomic systems by combining features from pre-trained graph neural networks (SCN, SchNet) with kernel mean embedding.
Transfer Learning with Affine Model Transformation
Shunya Minami (Institute of Statistical Mathematics), Ryo Yoshida (Institute of Statistical Mathematics)
CodeDomain AdaptationOptimizationTabular
🎯 What it does: This paper proposes a supervised transfer learning framework based on affine model transformation, which can simultaneously capture cross-domain commonalities and specific factors through affine mapping between source domain features and target domain outputs.
Transferable Adversarial Robustness for Categorical Data via Universal Robust Embeddings
Klim Kireev (École Polytechnique Fédérale de Lausanne), Nicolas Flammarion (École Polytechnique Fédérale de Lausanne)
CodeClassificationAnomaly DetectionOptimizationAdversarial AttackTabularFinance Related
🎯 What it does: A trainable robust neural network and a general robust embedding transferable to tree models are proposed for tabular data with categorical features, supporting a realistic threat model based on financial costs.
Transformer as a hippocampal memory consolidation model based on NMDAR-inspired nonlinearity
Dong-Kyum Kim (Institute for Basic Science), C. Justin Lee (Institute for Basic Science)
CodeTransformerSequential
🎯 What it does: A novel activation function based on the nonlinear dynamics of NMDA receptors is proposed, which is embedded in the feedforward network of the Transformer to simulate the hippocampal memory integration process and enhance long-term reference memory.
🎯 What it does: This paper proposes TPSR, which combines Monte Carlo Tree Search (MCTS) planning with pre-trained Transformers to introduce non-differentiable feedback on fitting accuracy and complexity in the generated program sequences during symbolic regression.
Transformers as Statisticians: Provable In-Context Learning with In-Context Algorithm Selection
Yu Bai (Salesforce Research), Song Mei (University of California Berkeley)
CodeTransformerTabular
🎯 What it does: This paper proves and implements that transformers can approximately execute various classic machine learning algorithms (ridge regression, Lasso, generalized linear models, two-layer neural network gradient descent) and an adaptive algorithm selection process in in-context learning (ICL), ultimately achieving near-Bayes optimal predictions.
Transformers learn to implement preconditioned gradient descent for in-context learning
Kwangjun Ahn (Massachusetts Institute of Technology), Suvrit Sra (Technical University of Munich)
CodeOptimizationTransformerTabular
🎯 What it does: The study investigates the training of Transformer on random linear regression instances, proving that a single-layer Transformer corresponds to a global optimum with one-step preconditioned gradient descent, while deeper layers correspond to multi-step preconditioned gradient methods.
Yuankai Luo (Beihang University), Lei Shi (Beihang University)
CodeGraph Neural NetworkTransformerGraph
🎯 What it does: An improved Transformer architecture for directed acyclic graphs (DAGs) is proposed, enabling it to capture partial sequential relationships of DAGs.
🎯 What it does: This paper proposes a Transformer with Hierarchical Prompting (TransHP), which introduces a hierarchical prompting mechanism in visual Transformers. It first predicts coarse categories through intermediate layers and then injects the corresponding prompt tokens into subsequent layers to enhance fine category classification performance.
Transition-constant Normalization for Image Enhancement
Jie Huang (University of Science and Technology of China), Feng Zhao (University of Science and Technology of China)
CodeRestorationImage
🎯 What it does: A new normalization operation called Transition-constant Normalization (TCN) is proposed for image enhancement tasks, balancing illumination consistency modeling and the reversibility of information transfer.
Transitivity Recovering Decompositions: Interpretable and Robust Fine-Grained Relationships
Abhra Chaudhuri (University of Exeter), Anjan Dutta (The Alan Turing Institute)
CodeClassificationExplainability and InterpretabilityGraph Neural NetworkTransformerImageAgriculture Related
🎯 What it does: An interpretable fine-grained visual classification method called TRD has been developed, which deconstructs abstract relational representations into graph structures and learns directly in the graph space, achieving complete interpretability at both the instance and category levels.
Tree of Thoughts: Deliberate Problem Solving with Large Language Models
Shunyu Yao (Princeton University), Karthik R Narasimhan
CodeTransformerLarge Language ModelPrompt EngineeringTextChain-of-Thought
🎯 What it does: Proposes the Tree of Thoughts (ToT) framework, which extends the reasoning of large language models from the word level to multi-path exploration and self-evaluation at the 'thought' level through tree search, in order to enhance the problem-solving ability for complex issues.
🎯 What it does: This paper proposes TreeVAE, a generative hierarchical clustering model that achieves hierarchical partitioning of samples and generates corresponding cluster samples by learning tree-structured posterior distributions in an unsupervised manner.
Tree-Rings Watermarks: Invisible Fingerprints for Diffusion Images
Yuxin Wen (University of Maryland), Tom Goldstein (University of Maryland)
CodeGenerationData SynthesisDiffusion modelImage
🎯 What it does: During the sampling process of the diffusion model, an 'tree ring' pattern is embedded in the Fourier domain of the initial noise vector, achieving an invisible watermark on the generated content without altering the generated images;
Triangulation Residual Loss for Data-efficient 3D Pose Estimation
Jiachen Zhao (Tsinghua University), Qionghai Dai (Tsinghua University)
CodePose EstimationImage
🎯 What it does: A Triangulation Residual Loss (TR loss) is proposed for multi-view 3D pose estimation, achieving unsupervised global geometric consistency training by minimizing the distance between 3D estimated points and rays from all views.
TriRE: A Multi-Mechanism Learning Paradigm for Continual Knowledge Retention and Promotion
Preetha Vijayan (NavInfo Europe), Elahe Arani (Eindhoven University of Technology)
CodeImage
🎯 What it does: A three-stage continual learning paradigm called TriRE is proposed, utilizing three mechanisms: retain, revise, and rewind to achieve the preservation, revision, and reshaping of task knowledge.
TrojLLM: A Black-box Trojan Prompt Attack on Large Language Models
Jiaqi Xue (University of Central Florida), Qian Lou (University of Central Florida)
CodeClassificationAdversarial AttackTransformerLarge Language ModelReinforcement LearningPrompt EngineeringText
🎯 What it does: This paper proposes a black-box backdoor trigger attack framework for large language model (LLM) APIs called TrojLLM, which can automatically generate general and covert trigger words and embed them into discrete prompts, leading to malicious manipulation of LLM outputs.
Truncated Affinity Maximization: One-class Homophily Modeling for Graph Anomaly Detection
Hezhe Qiao (Singapore Management University), Guansong Pang (Singapore Management University)
CodeAnomaly DetectionGraph Neural NetworkGraph
🎯 What it does: In unsupervised graph anomaly detection, a local node affinity anomaly scoring based on one-class homophily is proposed, and specialized node representations are learned through the Truncated Affinity Maximization (TAM) method, significantly improving anomaly detection performance.
Trust Region-Based Safe Distributional Reinforcement Learning for Multiple Constraints
Dohyeong Kim (Seoul National University), Songhwai Oh (Seoul National University)
CodeOptimizationSafty and PrivacyRobotic IntelligenceReinforcement Learning
🎯 What it does: A trust-region based safe distributed actor-critic algorithm (SDAC) is proposed, capable of training robots in multi-constraint and risk-averse reinforcement learning environments.
Tuning Multi-mode Token-level Prompt Alignment across Modalities
Dongsheng Wang (Xidian University), Hanwang Zhang (Nanyang Technological University)
CodeClassificationDomain AdaptationTransformerPrompt EngineeringVision Language ModelImageMultimodality
🎯 What it does: This paper proposes a method for simultaneously learning multi-modal visual and textual prompts in a multi-modal vision-language model, achieving token-level alignment through hierarchical optimal transport (OT) to enhance the performance of few-shot image classification.
Two Sides of The Same Coin: Bridging Deep Equilibrium Models and Neural ODEs via Homotopy Continuation
Shutong Ding (ShanghaiTech University), Ye Shi (ShanghaiTech University)
CodeClassificationExplainability and InterpretabilityComputational EfficiencyImageOrdinary Differential Equation
🎯 What it does: The HomoODE model is proposed, unifying Deep Equilibrium Models (DEQs) and Neural ODEs through the same theory (homotopy continuation method), and based on this, new implicit networks are constructed; at the same time, a strategy of sharing learnable initial points is proposed to accelerate the solving process, which explains the effects of Augmented Neural ODE.
Two-Stage Predict+Optimize for MILPs with Unknown Parameters in Constraints
Xinyi HU, Jimmy H.M. Lee (University of Wisconsin-Madison)
CodeOptimizationTabular
🎯 What it does: A Two-Stage Predict+Optimize framework is proposed to address mixed-integer linear programming (MILP) problems with unknown parameters in constraints, along with a general end-to-end gradient training algorithm.
UltraRE: Enhancing RecEraser for Recommendation Unlearning via Error Decomposition
Yuyuan Li (Zhejiang University), Jun Wang (OPPO Research Institute)
CodeRecommendation SystemTabular
🎯 What it does: This paper proposes a new recommendation system machine learning-free framework called UltraRE, aimed at improving model utility and learning-free efficiency while ensuring complete learning-free integrity.
🎯 What it does: A framework for fine-grained reward maximization (FGRM) is proposed to fine-tune a pre-trained evidence learning segmentation model using reinforcement learning, aiming for reliable uncertainty estimation in safety-critical scenarios.
🎯 What it does: This paper proposes an Uncertainty-Aware Inverse Propensity Score (UIPS) estimator to improve policy evaluation and optimization in offline policy learning with unknown logging policies.
Uncovering and Quantifying Social Biases in Code Generation
Yan Liu (Microsoft Research), Tsung-Yi Ho (Peking University)
CodeGenerationAI Code AssistantTransformerLarge Language ModelPrompt EngineeringText
🎯 What it does: This paper studies the issue of social bias in pre-trained code generation models, proposing a new code prompt template to activate model biases and quantifying the generated code for evaluation.
🎯 What it does: This paper proposes a prototype knowledge-based supervised method that improves the grouped token mechanism in ViT to achieve weakly supervised vocabulary semantic segmentation.
Understanding and Addressing the Pitfalls of Bisimulation-based Representations in Offline Reinforcement Learning
Hongyu Zang (Beijing Institute of Technology), Romain Laroche (Wayve)
CodeReinforcement LearningTabular
🎯 What it does: Research and improve the bisimulation-based state representation method in offline reinforcement learning to address the issues of estimation failure and feature collapse caused by missing offline data.
🎯 What it does: Theoretical proof is presented that ensemble adversarial defense outperforms single models, and the iGAT method is proposed to enhance ensemble robustness through global adversarial sample allocation and regularization.