Regularizes attention scores in vision transformers using bootstrapping.
problem Noisy and diffused attention maps in ViT limit interpretability.
method Statistical learning techniques, bootstrapping of attention scores.
result Improves shrinkage and sparsity of attention scores.
The paper quantifies the regularity of attention operations.
problem Quantifying the regularity of attention operations.
method Proposes a new mathematical framework using measure theory and integral operators.
result Proves attention operation is Lipschitz continuous on compact domains and provides an estimate of its Lipschitz constant.
Paper introduces regularization for multi-head attention to spot keywords.
problem Redundancy in multi-head attention leads to lack of rich information.
method Regularization technique to enforce orthogonality between attention heads.
result Significant improvement in keyword spotting performance.
The p-Laplacian Transformer improves transformer models by assigning higher attention weights to tokens in close proximity.
problem The self-attention mechanism in transformers does not effectively distinguish attention weights between tokens in close and non-close proximity.
method Proposes a novel class of transformers, p-Laplacian Transformers, that use p-Laplacian regularization to assign higher attention weights to tokens in close proximity. result Empirically demonstrates that p-Laplacian Transformers outperform baseline transformers on various benchmark datasets.
Transformers with linear space and time complexity for accurate attention estimation.
problem Efficiently estimating attention in large-scale tasks without relying on priors.
method Performers use Fast Attention Via positive Orthogonal Random features (FAVOR+) for linear approximation of softmax attention.
result Performers achieve competitive results on various tasks, demonstrating the effectiveness of their attention-learning approach.
The paper investigates polynomial alternatives to softmax in transformer models.
problem The effectiveness of softmax attention in transformers is questioned.
method The authors explore polynomial activations as alternatives to softmax, focusing on their ability to regularize the attention matrix.
result Certain polynomials can serve as effective substitutes for softmax in transformer applications, achieving strong performance.
Inspired by the adaptation phenomenon of neuronal firing, we propose the regularity normalization (RN) as an unsupervised attention mechanism (UAM) which computes the statistical regularity in the implicit space of neural networks under the Minimum Description Length (MDL) principle. Treating the neural network optimiz…
Researchers improve transformer networks' optimization and understanding.
problem Improving the understanding and optimization of transformer networks.
method Introducing a convex alternative to the self-attention mechanism and reformulating the training problem as a convex optimization problem.
result Revealed an implicit regularization mechanism that promotes sparsity across tokens.
Novel regularization for Vision Transformers improves model generalization and sparsity.
problem Improving generalization and sparsity in Vision Transformers.
method Likelihood-guided variational Ising-based regularization.
result Improved generalization and sparsity in Vision Transformers.
GCL-LRR improves node classification in noisy graphs.
problem Noise in real-world graph data impairs GNNs' effectiveness.
method Two-stage transductive learning with low-rank regularization and attention.
result Improved node classification performance in noisy graphs.
3D Axial-Attention improves lung nodule classification accuracy.
problem Limited 3D attention in existing methods.
method Proposes 3D Axial-Attention network with 3D positional encoding.
result 3D Axial-Attention achieves state-of-the-art performance.
Gradient flow on softmax attention minimizes nuclear norm of weight matrices.
problem Classification with separate key and query weight matrices.
method Gradient flow on exponential loss, separability assumption, reparameterization, approximate KKT conditions.
result Gradient flow implicitly minimizes nuclear norm of weight matrices, contrasting with Frobenius norm minimization.
Machine learning models that can exploit the inherent structure in data have gained prominence. In particular, there is a surge in deep learning solutions for graph-structured data, due to its wide-spread applicability in several fields. Graph attention networks (GAT), a recent addition to the broad class of feature le…
GOAT improves attention mechanisms by learning better priors.
problem Standard attention mechanisms use a naive uniform prior, limiting flexibility and generalization.
method GOAT introduces a trainable, continuous prior that replaces the uniform assumption, maintaining compatibility with optimized kernels.
result GOAT avoids representational trade-offs and learns an extrapolatable prior that combines positional flexibility with length generalization.
Improves dialogue response model interpretability using attention and regularization.
problem Improving interpretability of dual encoder models for dialogue response suggestions.
method Integrates attention mechanism and novel regularization loss to emphasize important words.
result Improves model accuracy and interpretability compared to existing methods.
Aligns attention distributions for improved accuracy and robustness.
problem Improving the accuracy and robustness of neural networks using attention mechanisms.
method Alignment attention that encourages key and query distributions to match within each head.
result Alignment attention leads to better accuracy, uncertainty estimation, and robustness across various tasks.
Adversarial training has been successfully applied to build robust models at a certain cost. While the robustness of a model increases, the standard classification accuracy declines. This phenomenon is suggested to be an inherent trade-off. We propose a model that employs feature prioritization by a nonlinear attention…
Transformers model contextual relations using probabilistic measures, revealing their expressive power.
problem Lack of clear understanding of Transformer's ability to model contextual relations.
method Introduced a measure-theoretic framework connecting softmax attention and entropy-regularized optimal transport.
result Transformer architectures can approximate arbitrary contextual relations, and the choice of normalization affects how these relations are represented.
We extend the Newlander-Nirenberg theorem to manifolds with almost complex structures that have somewhat less than Lipschitz regularity. We also discuss the regularity of local holomorphic coordinates in the integrable case, with particular attention to Lipschitz almost complex structures.
Proposes a time-aware attention model for CTR prediction.
problem Lack of temporal signals in existing CTR prediction models.
method Time-aware attention model with absolute and relative temporal signals, regularized adversarial sampling.
result Significantly improves CTR prediction performance.
Minimum attention improves reinforcement learning performance in high-dimensional dynamics.
problem Improving reinforcement learning performance in high-dimensional nonlinear dynamics.
method Applying minimum attention as a regularization technique in reinforcement learning, including model-based and model-free approaches.
result Minimum attention outperforms state-of-the-art algorithms in few-shot adaptation and variance reduction.
Bayesian attention modules improve model interpretability and performance.
problem Deterministic attention modules limit model interpretability and optimization.
method Proposes a scalable stochastic attention module using simplex-constrained distributions and Bayesian learning.
result Consistent improvements over baselines in various attention-based models.
The Generative Adversarial Network (GAN) has recently been applied to generate synthetic images from text. Despite significant advances, most current state-of-the-art algorithms are regular-grid region based; when attention is used, it is mainly applied between individual regular-grid regions and a word. These approach…
SGATs learn sparse attention coefficients to improve graph learning tasks on large, noisy graphs.
problem Overfitting and noisy edges in GNNs on large, noisy graphs.
method Sparse Graph Attention Networks (SGATs) learn sparse attention coefficients under L0-norm regularization. result SGATs can remove 50%-80% edges from large graphs while maintaining similar classification accuracies.
Modern neural networks are often augmented with an attention mechanism, which tells the network where to focus within the input. We propose in this paper a new framework for sparse and structured attention, building upon a smoothed max operator. We show that the gradient of this operator defines a mapping from real val…
This study analyzes how one-layer transformers learn regular language recognition tasks.
problem Understanding how one-layer transformers solve regular language recognition tasks like even pairs and parity check.
method Theoretical analysis of training dynamics and gradient descent for a one-layer transformer.
result A one-layer transformer can solve even pairs directly but needs CoT for parity check. Training phases show rapid growth in attention layer followed by logarithmic growth in linear layer.
GISST interprets GNNs by combining attention and sparsity for graph structure and node feature importance.
problem Lack of joint consideration of graph structure and node features in GNN interpretation.
method Model-agnostic framework using attention mechanism and sparsity regularization.
result GISST achieves superior node feature and edge explanation precision in synthetic and real-world datasets.
Gradient descent converges geometrically to optimal self-attention parameters.
problem Training softmax self-attention layers for linear regression.
method Structure-aware gradient descent with preconditioner and regularizer.
result Gradient descent converges geometrically to global minima.
Defines new metric space sections with Ahlfors-David regularity.
problem Defining and analyzing new types of sections in metric spaces.
method Introducing intrinsically quasi-symmetric sections and proving their Ahlfors-David regularity.
result Proves Ahlfors-David regularity for intrinsically quasi-symmetric sections.
Transformers can solve complex filtering problems for non-Gaussian signals.
problem Non-linear and non-Markovian filtering problems for conditionally Gaussian signals.
method Continuous-time transformer models called filterformers.
result Filterformers can approximate the conditional law of non-Markovian and conditionally Gaussian signal processes.
ETC improves Transformer models for long and structured inputs.
problem Scaling input length and encoding structured inputs in Transformers.
method Introduces global-local attention, relative position encodings, and CPC pre-training.
result Achieves state-of-the-art results on four natural language datasets.
We show that, for a closed orientable n-manifold, with n not congruent to 3 modulo 4, the existence of a CR-regular embedding into complex (n-1)-space ensures the existence of a totally real embedding into complex n-space. This implies that a closed orientable (4k+1)-manifold with non-vanishing Kervaire semi-characteri…
Inspired by the observation that humans are able to process videos efficiently by only paying attention where and when it is needed, we propose an interpretable and easy plug-in spatial-temporal attention mechanism for video action recognition. For spatial attention, we learn a saliency mask to allow the model to focus…
Variational methods for revealing visual concepts learned by convolutional neural networks have gained significant attention during the last years. Being based on noisy gradients obtained via back-propagation such methods require the application of regularization strategies. We present a mathematical framework unifying…
A new feature selection method using attention for neural networks.
problem Feature selection for neural networks with a budget constraint.
method Sequential Attention: greedy forward selection with attention weights.
result Achieves state-of-the-art results for neural networks.
Transfer learning through fine-tuning a pre-trained neural network with an extremely large dataset, such as ImageNet, can significantly accelerate training while the accuracy is frequently bottlenecked by the limited dataset size of the new target task. To solve the problem, some regularization methods, constraining th…
Improved hybrid acoustic model using interleaved self-attention and convolution.
problem Limited application of transformer in hybrid acoustic models.
method Proposed a model structure with interleaved self-attention and 1D convolution.
result Competitive recognition results on Librispeech dataset.
A new spectrum attention mechanism improves time series classification.
problem Improving robustness and classification accuracy in time series classification.
method Proposes a spectrum attention mechanism (SAM) to filter and highlight important frequency components, using L1 regularization and a tumbling window for segmentation.
result Experimental results show that the proposed SSAM method produces better feature representations and improves classification accuracy.
BERT-based architectures currently give state-of-the-art performance on many NLP tasks, but little is known about the exact mechanisms that contribute to its success. In the current work, we focus on the interpretation of self-attention, which is one of the fundamental underlying components of BERT. Using a subset of G…
New Performer model tackles long-sequence protein modeling.
problem Challenges of training complex Transformer models for long sequences.
method Linearly scalable long-context Transformer architecture, Performer.
result Performer provides strong theoretical guarantees and is effective for protein sequence modeling.
Sparse attention model reduces long-context inference time with exponential accuracy guarantees.
problem Efficiently processing long-context queries in large language models.
method Formalizes attention as a projection onto key vectors, analyzes entropic relaxation, and introduces Vashista Sparse Attention.
result Sparse attention concentrates on a constant-size active face, leading to exponential decay of inactive tokens' mass and linear scaling of active face error.
MCSAE improves speaker embedding by focusing on both high- and low-level features.
problem Reduced effect of low-level features in speaker embedding encoding.
method Masked cross self-attentive encoding using ResNet with multi-layer aggregation and random masking regularization.
result Improved speaker embedding with equal error rate of 2.63% and minimum detection cost function of 0.1453.
This work relaxes energy constraints in self-attention layers for a more general analysis.
problem Understanding inherent biases and dynamics in self-attention layers without energy functions.
method Dynamical systems analysis and Jacobian matrix examination.
result Normalized dynamics are close to a critical state, indicating high inference performance.
A new model detects and localizes anomalies in multivariate time series data.
problem Anomaly diagnosis in multivariate time series data, especially localization.
method Attention Low-Rank Transformer (ALoRa-T) with low-rank regularization and Attention Low-Rank score.
result The proposed method significantly outperforms state-of-the-art methods in anomaly detection and localization.
Sparse Hopfield model improves memory retrieval with fewer connections.
problem Memory retrieval efficiency with fewer connections.
method Sparse extension of Hopfield model, derived from sparse entropic regularizer.
result Sparse Hopfield model achieves tighter error bounds and better performance.
Network embedding represents nodes in a continuous vector space and preserves structure information from the Network. Existing methods usually adopt a "one-size-fits-all" approach when concerning multi-scale structure information, such as first- and second-order proximity of nodes, ignoring the fact that different scal…
Transformer models align words through attention weights, closely approximating Optimal Transport.
problem Understanding the internal mechanism of transformer models in language processing.
method Empirical evidence and theoretical analysis of attention weights and their relation to Optimal Transport.
result Transformer models can simulate gradient descent on the dual of entropy-regularized OT problem, providing a theoretical foundation for token alignment.
Over the past few years, trace regression models have received considerable attention in the context of matrix completion, quantum state tomography, and compressed sensing. Estimation of the underlying matrix from regularization-based approaches promoting low-rankedness, notably nuclear norm regularization, have enjoye…