Deep Learning

Con4m, a novel consistency learning framework, leverages contextual information to effectively classify segmented time series with inconsistent boundary labels and varying durations of classes, signif…

Boosting neural network prediction reliability, this research ingeniously combines statistical depth and Fermat distance for superior uncertainty quantification, eliminating the need for distributiona…

Collaborative Refining for Learning from Inaccurate Labels (CRL) refines data using annotator agreement, improving model accuracy with noisy labels.

CODA: A novel modeling scheme tackles subpopulation shifts in machine learning by disentangling spurious correlations, augmenting data strategically, and using reweighted consistency loss for improved…

Boosting in-distribution generalization is achieved by strategically altering the training data distribution to reduce simplicity bias and promote uniform feature learning.

This paper proposes cardinality-aware top-k classification, improving accuracy and efficiency by dynamically adjusting prediction set sizes.

A novel graph-theoretic framework bridges OOD detection & generalization, offering theoretical error bounds and competitive empirical performance.

Geometric Diffusion Bridge (GDB) accurately predicts geometric state evolution in complex systems by leveraging a probabilistic approach and equivariant diffusion processes, surpassing existing deep l…

Researchers break the curse of dimensionality in structured density estimation using graph resilience, a novel graphical parameter that effectively reduces the sample complexity.

Boosting graph neural networks: Topology-Invariant Pooling (TIP) leverages persistent homology to enhance graph pooling, achieving consistent performance gains across diverse datasets.

Boosting conformal prediction intervals improves accuracy and precision by tailoring them to specific desired properties via machine learning.

Diversified Block Sparse Bayesian Learning (DivSBL) improves block sparse signal recovery by adapting to unknown block structures, enhancing accuracy and robustness over existing methods.

Researchers drastically sped up high-fidelity deep learning model extraction, improving efficiency by up to 14.8x and challenging previous assumptions on the extraction bottleneck.

Strong pre-tuned MLPs and meta-tuned default parameters for GBDTs and MLPs improve tabular data classification and regression.

BAN: a novel backdoor defense using adversarial neuron noise for efficient detection and mitigation.

B-ary Tree Push-Pull (BTPP) achieves linear speedup for distributed learning on heterogeneous data, significantly outperforming state-of-the-art methods with minimal communication.

Attraos: a novel long-term time series forecasting model leveraging chaos theory, significantly outperforms existing methods by utilizing attractor dynamics for efficient and accurate prediction.

AROMA: Attentive Reduced Order Model with Attention enhances PDE modeling with local neural fields, offering efficient processing of diverse geometries and superior performance in simulating 1D and 2D…

Evidential deep learning’s uncertainty quantification is unreliable; this paper reveals its limitations, proposes model uncertainty incorporation for improved performance.

Cross-Attention-only Time Series Transformer (CATS) outperforms existing models by removing self-attention, improving long-term forecasting accuracy, and reducing computational cost.