Dimension d = O(m^{-2} log n) nearly achieves the optimal margin m^rd(+∞, A) for retrieval embeddings, with matching lower bounds showing d = O(k log(n/k)) suffices and is necessary for m = Θ(k^{-1/2}) on k-sparse query matrices.
hub Canonical reference
International conference on machine learning , pages=
Canonical reference. 80% of citing Pith papers cite this work as background.
42
Pith papers citing it
Background
80% of classified citations
hub tools
citation-role summary
background 4
baseline 1
citation-polarity summary
representative citing papers
Weighted InfoNCE objectives realize specific target geometries in embedding space, with SupCon producing size-dependent inter-class similarities under imbalance while Soft SupCon and certain continuous variants preserve regular simplex or unique optima.
In generalized contrastive learning with imbalanced classes, optimal representations collapse to class means whose angular geometry is determined by class proportions via convex optimization, and extreme imbalance causes all minority classes to collapse to one vector.
A myopic MINMPC framework learns a value function offline via inverse optimization from expert data, allowing short horizons with near-optimal performance and strict integer feasibility online for hybrid systems.
POPO uses bounded importance sampling on positive rollouts and a siamese policy network to achieve implicit negative gradients and stable optimization, matching or exceeding GRPO on math benchmarks such as 36.67% on AIME 2025.
The paper proves statistical consistency of contrastive loss to optimal ranking via an AUC criterion and derives generalization bounds O(1/m + 1/sqrt(n)) for supervised and O(1/sqrt(m) + 1/sqrt(n)) for self-supervised CRL that explain benefits of large negative sets.
PPAT residualizes losses via a prediction-powered control variate inside LURE, yielding lower-variance unbiased risk estimates, tailored acquisition, and asymptotic CIs that cover with fewer labels.
FedLAB organizes multimodal graph knowledge into typed hierarchical codebooks for modality evidence, node semantics, and topology context via federated semantic barycenter pre-training, improving performance by up to 7.53% on benchmarks while enabling semantic traceability.
SemiPrune uses a small labeled subset and semi-supervised pseudo-labeling to enable supervised dataset pruning methods, achieving state-of-the-art results on domain-specific, image-corrupted, and long-tailed datasets.
EvoStruct integrates evolutionary priors from a protein language model with structural priors from an E(3)-equivariant GNN to raise amino acid recovery by 16% and diversity by 2.3x on CHIMERA-Bench while cutting perplexity 43%.
Di-COT is an unsupervised contrastive method that stochastically partitions time-series windows into overlapping sub-blocks to learn representations without augmentation, reporting SOTA results on classification and transfer tasks across multiple benchmarks while cutting training time.
Supervised classification reaches neural collapse by design via normalized prototype losses on the hypersphere, outperforming CE and SCL on ImageNet-1K and other benchmarks with faster convergence and better transfer.
PerturbedVAE disentangles perturbation-specific signals from invariant gene expression structure to recover causal representations and improve out-of-distribution prediction in single-cell perturbation modeling.
GCE-MIL is a backbone-agnostic wrapper that directly optimizes MIL evidence for sufficiency, necessity, and recoverability, yielding modest gains in Macro-F1 and C-index plus more faithful patch selection across many backbones and datasets.
Pilot study uses pretrained video encoder features from lung ultrasound to predict 30-day CHF readmission, finding lower-lung views and temporal differences most informative with top MLP F1 of 0.80.
WorldComp2D explicitly structures latent space geometry by object identity and spatial proximity via a proximity-dependent encoder and localizer, cutting parameters up to 4X and FLOPs 2.2X versus state-of-the-art lightweight models on facial landmark localization while staying real-time on CPU.
Steering Dark Triad features in an LLM increases exploitative and aggressive behavior while leaving strategic deception and cognitive empathy unchanged, indicating dissociable antisocial pathways.
A dual-purpose benchmark supplies two text-derived knowledge graphs and one expert reference graph on the same biomedical corpus to jointly measure construction method quality and GNN robustness via semi-supervised node classification.
LIVEditor-14B applies a new sparse attention method (ISA) that prunes context and uses query-sharpness routing to cut attention latency ~60% with no loss in editing quality on standard benchmarks.
Triplet constraints realizable in D-dimensional Euclidean space cannot be preserved above 50% accuracy by any embedding of dimension at most cD for constant c<1, with UGC-hardness preventing better polynomial-time solutions in any dimension.
Exploiting data symmetries boosts k-NN to select near-optimal low-noise subsets from noisy datasets, approaching Bayes-optimal performance in high dimensions, with learned representations aiding partial symmetry knowledge.
RePrompT uses recurrent prompt tuning to inject prior-visit latent states and cohort-derived population prompt tokens into LLMs, yielding better performance than pure EHR or pure LLM baselines on MIMIC clinical prediction tasks.
LeJEPA derives an optimal isotropic Gaussian target for embeddings and enforces it via sketched regularization to deliver scalable, heuristics-free self-supervised pretraining with 79% ImageNet linear accuracy on ViT-H/14.
Video Prediction Policy conditions robot action learning on future-frame predictions inside fine-tuned video diffusion models, yielding 18.6% relative gains on Calvin ABC-D and 31.6% higher real-world success rates.
citing papers explorer
-
Is Dimensionality a Barrier for Retrieval Models?
Dimension d = O(m^{-2} log n) nearly achieves the optimal margin m^rd(+∞, A) for retrieval embeddings, with matching lower bounds showing d = O(k log(n/k)) suffices and is necessary for m = Θ(k^{-1/2}) on k-sparse query matrices.
-
A Unified Geometric Framework for Weighted Contrastive Learning
Weighted InfoNCE objectives realize specific target geometries in embedding space, with SupCon producing size-dependent inter-class similarities under imbalance while Soft SupCon and certain continuous variants preserve regular simplex or unique optima.
-
Optimal Representations for Generalized Contrastive Learning with Imbalanced Datasets
In generalized contrastive learning with imbalanced classes, optimal representations collapse to class means whose angular geometry is determined by class proportions via convex optimization, and extreme imbalance causes all minority classes to collapse to one vector.
-
Learning myopic mixed-integer nonlinear model predictive control from expert demonstrations
A myopic MINMPC framework learns a value function offline via inverse optimization from expert data, allowing short horizons with near-optimal performance and strict integer feasibility online for hybrid systems.
-
Beyond Negative Rollouts: Positive-Only Policy Optimization with Implicit Negative Gradients
POPO uses bounded importance sampling on positive rollouts and a siamese policy network to achieve implicit negative gradients and stable optimization, matching or exceeding GRPO on math benchmarks such as 36.67% on AIME 2025.
-
Statistical Consistency and Generalization of Contrastive Representation Learning
The paper proves statistical consistency of contrastive loss to optimal ranking via an AUC criterion and derives generalization bounds O(1/m + 1/sqrt(n)) for supervised and O(1/sqrt(m) + 1/sqrt(n)) for self-supervised CRL that explain benefits of large negative sets.
-
Prediction-Powered Active Testing
PPAT residualizes losses via a prediction-powered control variate inside LURE, yielding lower-variance unbiased risk estimates, tailored acquisition, and asymptotic CIs that cover with fewer labels.
-
FedLAB: Traceable Semantic Codebooks for Federated Multimodal Graph Foundation Learning
FedLAB organizes multimodal graph knowledge into typed hierarchical codebooks for modality evidence, node semantics, and topology context via federated semantic barycenter pre-training, improving performance by up to 7.53% on benchmarks while enabling semantic traceability.
-
Label-Efficient Dataset Pruning via Semi-Supervised Pseudo-Labeling
SemiPrune uses a small labeled subset and semi-supervised pseudo-labeling to enable supervised dataset pruning methods, achieving state-of-the-art results on domain-specific, image-corrupted, and long-tailed datasets.
-
EvoStruct: Bridging Evolutionary and Structural Priors for Antibody CDR Design via Protein Language Model Adaptation
EvoStruct integrates evolutionary priors from a protein language model with structural priors from an E(3)-equivariant GNN to raise amino acid recovery by 16% and diversity by 2.3x on CHIMERA-Bench while cutting perplexity 43%.
-
Divide and Contrast: Learning Robust Temporal Features without Augmentation
Di-COT is an unsupervised contrastive method that stochastically partitions time-series windows into overlapping sub-blocks to learn representations without augmentation, reporting SOTA results on classification and transfer tasks across multiple benchmarks while cutting training time.
-
Neural Collapse by Design: Learning Class Prototypes on the Hypersphere
Supervised classification reaches neural collapse by design via normalized prototype losses on the hypersphere, outperforming CE and SCL on ImageNet-1K and other benchmarks with faster convergence and better transfer.
-
What Makes a Representation Good for Single-Cell Perturbation Prediction?
PerturbedVAE disentangles perturbation-specific signals from invariant gene expression structure to recover causal representations and improve out-of-distribution prediction in single-cell perturbation modeling.
-
GCE-MIL: Faithful and Recoverable Evidence for Multiple Instance Learning in Whole-Slide Imaging
GCE-MIL is a backbone-agnostic wrapper that directly optimizes MIL evidence for sufficiency, necessity, and recoverability, yielding modest gains in Macro-F1 and C-index plus more faithful patch selection across many backbones and datasets.
-
Prognostic Value of Lung Ultrasound Biomarkers for Readmission Risk in Congestive Heart Failure: A Pilot Data-Driven Analysis
Pilot study uses pretrained video encoder features from lung ultrasound to predict 30-day CHF readmission, finding lower-lung views and temporal differences most informative with top MLP F1 of 0.80.
-
WorldComp2D: Spatio-semantic Representations of Object Identity and Location from Local Views
WorldComp2D explicitly structures latent space geometry by object identity and spatial proximity via a proximity-dependent encoder and localizer, cutting parameters up to 4X and FLOPs 2.2X versus state-of-the-art lightweight models on facial landmark localization while staying real-time on CPU.
-
Exploitation Without Deception: Dark Triad Feature Steering Reveals Separable Antisocial Circuits in Language Models
Steering Dark Triad features in an LLM increases exploitative and aggressive behavior while leaving strategic deception and cognitive empathy unchanged, indicating dissociable antisocial pathways.
-
A Unified Benchmark for Evaluating Knowledge Graph Construction Methods and Graph Neural Networks
A dual-purpose benchmark supplies two text-derived knowledge graphs and one expert reference graph on the same biomedical corpus to jointly measure construction method quality and GNN robustness via semi-supervised node classification.
-
LIVEditor-14B: Lightning Unified Video Editing via In-Context Sparse Attention
LIVEditor-14B applies a new sparse attention method (ISA) that prunes context and uses query-sharpness routing to cut attention latency ~60% with no loss in editing quality on standard benchmarks.
-
Provable Accuracy Collapse in Embedding-Based Representations under Dimensionality Mismatch
Triplet constraints realizable in D-dimensional Euclidean space cannot be preserved above 50% accuracy by any embedding of dimension at most cD for constant c<1, with UGC-hardness preventing better polynomial-time solutions in any dimension.
-
Leveraging Data Symmetries to Select an Optimal Subset of Training Data under Label Noise
Exploiting data symmetries boosts k-NN to select near-optimal low-noise subsets from noisy datasets, approaching Bayes-optimal performance in high dimensions, with learned representations aiding partial symmetry knowledge.
-
RePrompT: Recurrent Prompt Tuning for Integrating Structured EHR Encoders with Large Language Models
RePrompT uses recurrent prompt tuning to inject prior-visit latent states and cohort-derived population prompt tokens into LLMs, yielding better performance than pure EHR or pure LLM baselines on MIMIC clinical prediction tasks.
-
LeJEPA: Provable and Scalable Self-Supervised Learning Without the Heuristics
LeJEPA derives an optimal isotropic Gaussian target for embeddings and enforces it via sketched regularization to deliver scalable, heuristics-free self-supervised pretraining with 79% ImageNet linear accuracy on ViT-H/14.
-
Video Prediction Policy: A Generalist Robot Policy with Predictive Visual Representations
Video Prediction Policy conditions robot action learning on future-frame predictions inside fine-tuned video diffusion models, yielding 18.6% relative gains on Calvin ABC-D and 31.6% higher real-world success rates.
-
MetaMorph: Multimodal Understanding and Generation via Instruction Tuning
VPiT enables pretrained LLMs to perform both visual understanding and generation by predicting discrete text tokens and continuous visual tokens, with understanding data proving more effective than generation-specific data.
-
Smaug: Fixing Failure Modes of Preference Optimisation with DPO-Positive
DPOP is a new loss function that prevents DPO from lowering preferred response likelihoods and outperforms standard DPO on diverse datasets, MT-Bench, and enables Smaug-72B to exceed 80% on the Open LLM Leaderboard.
-
Atlas: Few-shot Learning with Retrieval Augmented Language Models
Atlas reaches over 42% accuracy on Natural Questions with only 64 examples, outperforming a 540B-parameter model by 3% with 50x fewer parameters.
-
Unsupervised Dense Information Retrieval with Contrastive Learning
Contrastive learning trains unsupervised dense retrievers that beat BM25 on most BEIR datasets and support cross-lingual retrieval across scripts.
-
Zeus: Towards Tuning-Free Foundation Model for Time Series Analysis
Zeus proposes a multi-scale Transformer with point-wise tokenization and Multi-Objective Temporal Masking to enable tuning-free performance on forecasting, interpolation, and other time series tasks.
-
Multi-Scale Contrastive Attention for Light-Curve Representation Learning
Astra-CLR is a multi-filter time-series Transformer pre-trained via contrastive learning on 2.1 million ZTF light curves that achieves 0.70 accuracy classifying 12 variability classes, rising to 0.77 with partial fine-tuning.
-
PACD-Net: Pseudo-Augmented Contrastive Distillation for Glycemic Control Estimation from SMBG
PACD-Net uses pseudo-augmented contrastive distillation with a hybrid Swin Transformer-CNN backbone to estimate TAR, TIR, and TBR from sparse SMBG data and outperforms prior methods in accuracy and stability under sparse conditions.
-
Temporal Aware Pruning for Efficient Diffusion-based Video Generation
TAPE applies temporal-aware token pruning with smoothing, reselection, and timestep scheduling to speed up video diffusion models while preserving visual fidelity and coherence.
-
Beyond Instance-Level Self-Supervision in 3D Multi-Modal Medical Imaging
A self-supervised approach uses consistent spatial relationships of anatomical structures across patients to improve 3D multi-modal medical image representations, yielding modest gains on segmentation and classification tasks.
-
HEDP: A Hybrid Energy-Distance Prompt-based Framework for Domain Incremental Learning
HEDP uses energy regularization inspired by Helmholtz free energy plus hybrid energy-distance weighting in prompts to improve domain selection and achieve a 2.57% accuracy gain on benchmarks like CORe50 while mitigating catastrophic forgetting.
-
Structured Diffusion Bridges: Inductive Bias for Denoising Diffusion Bridges
A structured diffusion bridge method achieves near fully-paired modality translation quality using alignment constraints even in unpaired or semi-paired regimes.
-
Learning Fingerprints for Medical Time Series with Redundancy-Constrained Information Maximization
A self-supervised method learns a fixed set of disentangled fingerprint tokens from medical time series by combining reconstruction loss with a total coding rate diversity penalty, framed as a disentangled rate-distortion problem.
-
Learning Invariant Modality Representation for Robust Multimodal Learning from a Causal Inference Perspective
CmIR uses causal inference to separate invariant causal representations from spurious ones in multimodal data, improving generalization under distribution shifts and noise via invariance, mutual information, and reconstruction constraints.
-
Detecting Breast Carcinoma Metastasis on Whole-Slide Images by Partially Subsampled Multiple Instance Learning
A Gaussian mixture MIL framework with partially subsampled instances improves metastasis prediction accuracy on breast cancer whole-slide images over prior methods.
-
The Platonic Representation Hypothesis
Representations learned by large AI models are converging toward a shared statistical model of reality.
-
Retrieve Only Relevant Tables Whether Few or Many: Adaptive Table Retrieval Method
An adaptive thresholding mechanism combined with sliding-window reranking retrieves a query-dependent number of tables from large corpora, improving retrieval and downstream text-to-SQL performance on Spider, BIRD, and Spider 2.0.
- ConTact: Contact-First Antibody CDR Design via Explicit Interface Reasoning
- Online Goal Recognition using Path Signature and Dynamic Time Warping