Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,657 papers · 148 categories

Trend · papers per month

95191286381 · Jun 202019922001200920172026
48 results for attention weights

Attention weights may not accurately highlight important parts due to combinatorial shortcuts.

problem Inaccurate interpretation of attention weights in models.
method Theoretical analysis and design of experiments to show combinatorial shortcuts. Proposed two methods to mitigate this issue.
result Proposed methods improve the interpretability of attention mechanisms.

This research integrates attention into XAI frameworks for better model explanations.

problem Improving the interpretability of transformer models.
method Developed two novel explanation methods: Shapley value decomposition and token-level directional derivatives.
result Attention weights can be meaningfully incorporated into XAI frameworks, enhancing transformer explainability.

Attention networks show promise for both vision and language tasks, by emphasizing relationships between constituent elements through weighting functions. Such elements could be regions in an image output by a region proposal network, or words in a sentence, represented by word embedding. Thus far the learning of atten…

2019-05-27abs ↗pdf ↗

ZeroS improves Transformers by adding negative weights, matching or beating softmax attention.

problem Limited performance of linear attention methods, especially in long context sequences.
method Proposes Zero-Sum Linear Attention (ZeroS) that removes the zero-order term and reweights zero-sum softmax residuals.
result ZeroS matches or exceeds standard softmax attention across various benchmarks, theoretically expanding representable functions.

Elliptical Attention improves transformer performance by focusing on contextually relevant features.

problem Transformer models suffer from representation collapse and are vulnerable to contaminated samples.
method Uses Mahalanobis distance to define hyper-ellipsoidal neighborhoods for attention weights.
result Elliptical Attention reduces representation collapse and enhances model robustness.

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.

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 pp-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.

As an important part of speech recognition technology, automatic speech keyword recognition has been intensively studied in recent years. Such technology becomes especially pivotal under situations with limited infrastructures and computational resources, such as voice command recognition in vehicles and robot interact…

2019-07-10abs ↗pdf ↗

Graph Attention Networks (GATs) are the state-of-the-art neural architecture for representation learning with graphs. GATs learn attention functions that assign weights to nodes so that different nodes have different influences in the feature aggregation steps. In practice, however, induced attention functions are pron…

2019-10-25abs ↗pdf ↗

Study shows attention-style models learn pairwise interactions efficiently.

problem Learning pairwise interactions in attention-style models.
method Proved minimax rate of convergence for learning pairwise interactions.
result Minimax rate is M2β2β+1M^{-\frac{2β}{2β+1}} independent of embedding dimension and token number.

Proposes new methods for interpreting document classification models.

problem Interpretation fragility of attention-based neural networks.
method Corpus-level and concept-based explanation methods using attention weights.
result Extracts semantically meaningful keywords and concepts for model predictions.

This paper recovers input data from transformer models using attention weights.

problem Recovering input data from transformer models for security and privacy concerns.
method Introducing an algorithm to minimize the loss function between expected and actual outputs of transformers.
result The algorithm successfully recovers input data from attention weights and outputs of transformers.

Study reveals how attention helps in signal recovery from sequence models using random matrix theory.

problem Signal recovery from sequence models with attention mechanisms.
method Analysis of sample covariance matrices constructed from pooled sequence representations with attention weights.
result Optimal attention weights maximize signal-to-noise ratio and improve signal recovery.

SpGAT learns graph representations using spectral attention for efficiency.

problem Efficiently capturing global graph patterns with minimal parameters.
method Introduces Spectral Graph Attention Network (SpGAT) using spectral domain attention mechanisms and a fast Chebychev approximation.
result SpGAT achieves better global pattern recognition with fewer parameters compared to GAT.

Although deep neural networks generally have fixed network structures, the concept of dynamic mechanism has drawn more and more attention in recent years. Attention mechanisms compute input-dependent dynamic attention weights for aggregating a sequence of hidden states. Dynamic network configuration in convolutional ne…

2019-12-01abs ↗pdf ↗

VALC provides concept-level interpretations of FLMs, overcoming word-level limitations.

problem Lack of higher-level structure interpretation in FLMs' attention weights.
method Formal definition of conceptual interpretation, variational Bayesian framework (VALC).
result VALC finds optimal language concepts for FLM predictions, providing concept-level interpretations.

A new method for feature fusion in U-Net decoders using difference-based gating.

problem Precise fusion of high-level semantics and low-level details in U-Net decoder reconstruction.
method Proposes two difference-based gating approaches: Feature-difference gating (FDG) and Entropy-difference gating (EDG).
result Both FDG and EDG methods outperform existing attention-based fusion methods, with EDG showing superior performance.

New analysis shows how cross-entropy training shapes attention in transformers.

problem Understanding how gradient-based learning creates the required internal geometry in transformers.
method Developed a first-order analysis of cross-entropy training effects on attention scores and values in a transformer attention head.
result Introduced an advantage-based routing law and responsibility-weighted update for attention scores and values, respectively.

Attention mechanisms have become ubiquitous in NLP. Recent architectures, notably the Transformer, learn powerful context-aware word representations through layered, multi-headed attention. The multiple heads learn diverse types of word relationships. However, with standard softmax attention, all attention heads are de…

2019-08-30abs ↗pdf ↗

This paper analyzes the interpolation error of nonlinear Attention compared to linear regression.

problem Understanding the interpolation error of nonlinear Attention in high-dimensional settings.
method Derives explicit expressions for mean-squared interpolation error using signal-plus-noise model and random matrix theory.
result Nonlinear Attention generally incurs a larger interpolation error than linear regression, but this gap can be reversed with structured signals.

Adversarial domain-invariant training (ADIT) proves to be effective in suppressing the effects of domain variability in acoustic modeling and has led to improved performance in automatic speech recognition (ASR). In ADIT, an auxiliary domain classifier takes in equally-weighted deep features from a deep neural network …

2019-04-28abs ↗pdf ↗

Transformers can emulate various algorithms by prompting, proving universality.

problem How to emulate algorithms using fixed-weight Transformers.
method Two modes of in-context algorithm emulation: task-specific and prompt-programmable. Constructing prompts that encode algorithm parameters into token representations.
result Fixed-weight Transformers can emulate a broad class of algorithms via prompts.

WildCat efficiently compresses neural network attention mechanisms.

problem Expensive quadratic runtime of attention mechanisms in neural networks.
method Uses a weighted coreset with a fast subsampling algorithm to approximate attention with near-linear time complexity.
result Approximates exact attention with super-polynomial error decay and near-linear runtime.

KATA improves associative recall by optimizing feature maps derived from nonnegative attention weights.

problem Linear attention's poor performance on associative recall tasks.
method Formulates attention recall as a spherical-packing problem and introduces Kernelized Linear Attention Activations (KATA).
result KATA features offer a favorable capacity-interference tradeoff, enabling efficient associative recall.

The study reveals the spectral structure of attention layers and its implications for generalization.

problem Understanding the spectral structure and generalization of trained attention layers.
method Empirical risk minimization in a single-head tied-attention layer, using random matrix theory, spin-glass theory, and approximate message passing.
result Exact high-dimensional characterization of training and test error, interpolation and recovery thresholds, and spectrum of the key and query matrices.

Attention-based models have recently shown great performance on a range of tasks, such as speech recognition, machine translation, and image captioning due to their ability to summarize relevant information that expands through the entire length of an input sequence. In this paper, we analyze the usage of attention mec…

2017-10-28abs ↗pdf ↗

We present a general architecture of deep differentiable forest and its sparse attention mechanism. The differentiable forest has the advantages of both trees and neural networks. Its structure is a simple binary tree, easy to use and understand. It has full differentiability and all variables are learnable parameters.…

2020-02-29abs ↗pdf ↗

HKT improves sequence processing with multi-scale attention and kernel analysis.

problem Processing sequences at multiple scales with efficient attention mechanisms.
method Trainable causal downsampling and convex weights for level-specific score matrices.
result HKT achieves consistent gains over standard attention across various tasks.

CLEAR learns causal graphs from attention in recommender systems to explain user behavior.

problem Understanding why specific recommendations are made in recommender systems.
method CLEAR learns session-specific causal graphs from attention in pre-trained neural recommenders, addressing latent confounders.
result CLEAR provides counterfactual explanations that are shorter and more effective than naive methods.

The paper identifies patterns in language model weights used for memorizing paragraphs.

problem Locating the specific mechanisms and weights used by language models to memorize paragraphs.
method Examined gradients and attention patterns in language models to identify memorized paragraphs.
result Gradients of memorized paragraphs have a distinguishable spatial pattern, and localized attention heads are involved in paragraph memorization.

A novel CAB-XDE framework predicts speculative stock prices with high accuracy.

problem Forecasting speculative stock prices in volatile markets.
method Customized attention BiLSTM with XGBoost, integrating attention mechanism and weight determination theory-error reciprocal method.
result Empirically validated with MAPE of 0.0037, MAE of 84.40, and RMSE of 106.14.