Representation Learning

Contrastive learning enables efficient probabilistic inference in high-dimensional time series by creating Gaussian representations that form a Gauss-Markov chain, allowing for closed-form solutions t…

Researchers can now pinpoint the sources of data shifts in complex linear causal systems using a new algorithm, even with limited perfect interventions, opening exciting possibilities for causal disco…

Probabilistic Slot Attention achieves identifiable object-centric representations without supervision, advancing systematic generalization in machine learning.

This paper presents a novel approach for embedding supervised models in hyperbolic space, linking loss functions to hyperbolic distances and introducing monotonic decision trees for unambiguous visual…

TAG-CF boosts collaborative filtering accuracy by up to 39.2% on cold users, using only a single message-passing step at test time, avoiding costly training-time computations.

HC-GAE: A novel hierarchical graph autoencoder combats over-smoothing by using hard node assignment to create isolated subgraphs, improving graph representation learning for classification.

GraphCroc, a novel graph autoencoder, leverages cross-correlation to accurately reconstruct complex graph structures, outperforming self-correlation-based methods.

Graphcodes efficiently summarize complex datasets’ topological properties using graph neural networks, enhancing machine learning accuracy.

GEM, a novel framework, uses a bidirectional graph and MLLMs to achieve fine-grained, relation-aware disentanglement in unsupervised representation learning, surpassing existing methods.

Reward-driven distortions in grid cell patterns are global, not local, preserving path integration while encoding environmental landmarks in spatial navigation.

Researchers derived tighter generalization bounds for label-specific representation learning (LSRL) methods, improving understanding of LSRL’s success and offering guidance for future algorithm develo…

GIN, a novel approximate Bayesian inference algorithm, efficiently handles nonlinear state-space models with high-dimensional, noisy observations by disentangling observation and dynamics. Achieving l…

This paper introduces a novel geometric approach to concept-based representation learning, provably recovering interpretable concepts from diverse data without strict causal assumptions or many interv…

Single-layer GANs learn data subspaces more effectively using multi-feature discriminators, enabling faster training and better feature representation than conventional methods.

Boost GNN graph classification accuracy by enforcing consistency in learned representations across layers using a novel loss function!

Pretrained neural networks surprisingly capture fundamental aspects of human cognition, enabling generalization in image-based learning tasks, as demonstrated by aligning neural network representation…

This paper presents a novel approach to deriving theoretically grounded disentanglement metrics by linking logical definitions to quantitative measures, offering strong theoretical guarantees and easi…

Hierarchical Distance Structural Encoding (HDSE) empowers graph transformers to better capture hierarchical graph structures, leading to improved performance in graph classification and regression tas…

Generative models learn hidden capabilities suddenly during training, which can be explained and predicted using a novel ‘concept space’ framework that analyzes learning dynamics and concept signal.

Discover optimal embedding dimensions for contrastive learning & k-NN using graph arboricity; achieve efficient model design & performance.