Model collapse occurs in structured interactive learning if and only if the directed interaction graph satisfies a specific topological condition, with finite-sample guarantees for linear regression and asymptotic results for M-estimators.
hub
Advances in neural information processing systems , volume=
20 Pith papers cite this work. Polarity classification is still indexing.
20
Pith papers citing it
hub tools
representative citing papers
A novel decoupled method for distributed saddle problems achieves optimal communication complexity via multi-stage residual norm minimization, with a matching lower bound and extension to variational inequalities.
A single-network fixed-point formulation for neural optimal transport eliminates adversarial min-max optimization and implicit differentiation while enforcing dual feasibility exactly.
A two-stage symbolic regression plus generative model framework recovers governing interaction terms and forcing in stochastic triad models while accurately predicting statistical moments up to order five.
ActDiff-VC achieves up to 64.6% bitrate reduction at matched NIQE and improves perceptual metrics like KID and FID by using content-adaptive keyframe selection and budget-aware sparse trajectory selection to condition a diffusion decoder for ultra-low-bitrate video reconstruction.
RetinaDiff uses phase correlation to create a motion-corrected physics prior and then applies a conditional diffusion model to reconstruct high-quality temporal LSCI from extremely limited frames.
Min-K% Prob detects pretraining data in LLMs by flagging outlier low-probability words in text, achieving 7.4% better performance than prior methods on the new WIKIMIA benchmark.
Introduces a fairness layer for deep learning models that guarantees output parity and an online primal-dual algorithm for aggregate fairness guarantees in streaming predictions with small batch sizes.
AnyAct generates editable human reenactments from character videos via conditional motion generation from transferable sparse local 2D articulated cues, with designs for human-only supervision and global-local decoupling.
DeconDTN-Toolkit simulates provenance shifts to expose ERM vulnerabilities and provides tools plus a robust OOD indicator for mitigating confounding by data provenance.
Predictive Bayesian inference posteriors concentrate onto a forward-model-dependent quantity and produce miscalibrated credible sets unless the predictive model contains the true data-generating process.
Denoising Recursion Models train multi-step noise reversal in looped transformers and outperform the prior Tiny Recursion Model on ARC-AGI.
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.
A single-objective rectified flow variant uses neural ODEs trained by regression to monotonically decrease a fixed convex transport cost while preserving marginal distributions.
ERPPO adds a DSA-based ambiguity estimator to MAPPO and switches between L1 and L2 entropy regularization to improve exploration and stability in non-stationary multi-dimensional observations.
LADS is a sampling method that keeps benign user generations statistically identical to the original model while forcing correlated samples across a distiller's multiple accounts, provably worsening their generalization via uniform convergence bounds.
TaTok is a theoretically grounded adaptive tokenization method that uses global tokens and cumulative conditional entropy filtering to reduce redundancy while improving reconstruction quality over fixed-rate patch tokenization.
A structured diffusion bridge method achieves near fully-paired modality translation quality using alignment constraints even in unpaired or semi-paired regimes.
Discriminator-informed resampling via normalizing flows reduces error in the EnGMF for low-ensemble regimes on the Ikeda map and Lorenz '63 system.
Adversarial optimal transport objectives train neural emulators with improved long-term statistical fidelity on chaotic systems.
citing papers explorer
-
When Does Model Collapse Occur in Structured Interactive Learning?
Model collapse occurs in structured interactive learning if and only if the directed interaction graph satisfies a specific topological condition, with finite-sample guarantees for linear regression and asymptotic results for M-estimators.
-
Efficient Gradient Methods for Distributed Saddle Problems
A novel decoupled method for distributed saddle problems achieves optimal communication complexity via multi-stage residual norm minimization, with a matching lower bound and extension to variational inequalities.
-
Fixed-Point Neural Optimal Transport without Implicit Differentiation
A single-network fixed-point formulation for neural optimal transport eliminates adversarial min-max optimization and implicit differentiation while enforcing dual feasibility exactly.
-
The finite expression method for turbulent dynamics with high-order moment recovery
A two-stage symbolic regression plus generative model framework recovers governing interaction terms and forcing in stochastic triad models while accurately predicting statistical moments up to order five.
-
Active Sampling for Ultra-Low-Bit-Rate Video Compression via Conditional Controlled Diffusion
ActDiff-VC achieves up to 64.6% bitrate reduction at matched NIQE and improves perceptual metrics like KID and FID by using content-adaptive keyframe selection and budget-aware sparse trajectory selection to condition a diffusion decoder for ultra-low-bitrate video reconstruction.
-
Physics-Informed Conditional Diffusion for Motion-Robust Retinal Temporal Laser Speckle Contrast Imaging
RetinaDiff uses phase correlation to create a motion-corrected physics prior and then applies a conditional diffusion model to reconstruct high-quality temporal LSCI from extremely limited frames.
-
Detecting Pretraining Data from Large Language Models
Min-K% Prob detects pretraining data in LLMs by flagging outlier low-probability words in text, achieving 7.4% better performance than prior methods on the new WIKIMIA benchmark.
-
Differentiable Optimization Layers for Guaranteed Fairness in Deep Learning
Introduces a fairness layer for deep learning models that guarantees output parity and an online primal-dual algorithm for aggregate fairness guarantees in streaming predictions with small batch sizes.
-
AnyAct: Towards Human Reenactment of Character Motion From Video
AnyAct generates editable human reenactments from character videos via conditional motion generation from transferable sparse local 2D articulated cues, with designs for human-only supervision and global-local decoupling.
-
DeconDTN-Toolkit: A Library for Evaluation and Enhancement of Robustness to Provenance Shift
DeconDTN-Toolkit simulates provenance shifts to expose ERM vulnerabilities and provides tools plus a robust OOD indicator for mitigating confounding by data provenance.
-
Concentration and Calibration in Predictive Bayesian Inference
Predictive Bayesian inference posteriors concentrate onto a forward-model-dependent quantity and produce miscalibrated credible sets unless the predictive model contains the true data-generating process.
-
One Step Forward and K Steps Back: Better Reasoning with Denoising Recursion Models
Denoising Recursion Models train multi-step noise reversal in looped transformers and outperform the prior Tiny Recursion Model on ARC-AGI.
-
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.
-
Rectified Flow: A Marginal Preserving Approach to Optimal Transport
A single-objective rectified flow variant uses neural ODEs trained by regression to monotonically decrease a fixed convex transport cost while preserving marginal distributions.
-
ERPPO: Entropy Regularization-based Proximal Policy Optimization
ERPPO adds a DSA-based ambiguity estimator to MAPPO and switches between L1 and L2 entropy regularization to improve exploration and stability in non-stationary multi-dimensional observations.
-
Lossless Anti-Distillation Sampling
LADS is a sampling method that keeps benign user generations statistically identical to the original model while forcing correlated samples across a distiller's multiple accounts, provably worsening their generalization via uniform convergence bounds.
-
Mutual Enhancement Between Global Tokens and Patch Tokens: From Theory to Practice
TaTok is a theoretically grounded adaptive tokenization method that uses global tokens and cumulative conditional entropy filtering to reduce redundancy while improving reconstruction quality over fixed-rate patch tokenization.
-
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 Discriminators for Resampling in the Ensemble Gaussian Mixture Filter through a Normalizing Flow Approach
Discriminator-informed resampling via normalizing flows reduces error in the EnGMF for low-ensemble regimes on the Ikeda map and Lorenz '63 system.
-
Learning to Emulate Chaos: Adversarial Optimal Transport Regularization
Adversarial optimal transport objectives train neural emulators with improved long-term statistical fidelity on chaotic systems.