Structured State-Space Duality connects SSMs to masked attention.
problem Connecting SSMs and attention mechanisms for efficient modeling.
method Formalizing and generalizing SSD from scalar-identity to diagonal state matrices.
result Diagonal SSMs match training complexity lower bounds and support richer dynamics.
New analysis shows how attention masks and LayerNorm prevent rank collapse in transformers.
problem Rank collapse in transformer models with increasing depth.
method General analysis of rank collapse under self-attention, considering attention masks and LayerNorm.
result Self-attention with LayerNorm can prevent rank collapse and maintain a rich set of equilibria.
This paper improves multichannel speech enhancement using complex ratio masking and channel-attention.
problem Limited performance of deep learning methods in multichannel speech enhancement.
method Introduces complex ratio masking and channel-attention mechanism inside a U-Net architecture.
result Demonstrates superior performance on the CHiME-3 dataset.
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.
A new method for graph-structured data improves transformer performance by incorporating topology.
problem Improving transformer performance on graph-structured data.
method Parameterizing topological masks as a learnable function of a weighted adjacency matrix, approximated with graph random features.
result Efficient masking algorithms provide strong performance gains for tasks on image and point cloud data.
Transformers' self-attention mechanism is mapped to a generalized Potts model.
problem Uncertainty in what type of data distribution self-attention can efficiently learn.
method Decouple word positions and embeddings, then show self-attention learns a generalized Potts model.
result Training self-attention is equivalent to solving the inverse Potts problem.
Motivated by the fact that most of the information relevant to the prediction of target tokens is drawn from the source sentence S=s1,…,sS, we propose truncating the target-side window used for computing self-attention by making an N-gram assumption. Experiments on WMT EnDe and EnFr data sets show that the…
Improved YOLOv5 model detects mask-wearing with enhanced accuracy.
problem Detecting mask-wearing in high traffic public places.
method Improved YOLOv5l with Multi-Head Attentional Self-Convolution, Swin Transformer Block, I-CBAM module, and enhanced feature fusion.
result 1.1% improvement in mAP(0.5) and 1.3% improvement in mAP(0.5:0.95) compared to YOLOv5l.
LatFormer improves geometric reasoning by incorporating lattice symmetry priors in attention mechanisms.
problem State-of-the-art models struggle with geometric reasoning tasks in the ARC and LARC datasets.
method Introduced LatFormer, a model that uses lattice symmetry priors in attention masks.
result LatFormer requires 2 orders of magnitude fewer data than standard attention mechanisms.
Proposes a proportional masking strategy for better tabular data imputation.
problem Heterogeneity of tabular data disrupts uniform random masking in MAEs.
method Computes missingness statistics, generates proportional masks, uses MLP token mixing.
result Proportional masking preserves missingness distribution, improves imputation performance.
Enhances Transformer models for multivariate time series with dataset-specific channel masks.
problem Capturing channel dependency in multivariate time series models.
method Introduces partial channel dependence and channel masks to refine channel dependency in Transformers.
result Improves model performance across diverse tasks and datasets.
Model detects electricity theft with high accuracy.
problem Detecting electricity theft on imbalanced datasets.
method Multi-head self-attention mechanism with dilated convolutions and binary mask.
result Achieved AUC of 0.926, improving previous work by 17%.
Transformers exhibit abrupt learning in matrix completion tasks.
problem Understanding abrupt learning in Transformers for matrix completion.
method Formulated matrix completion as MLM task, trained BERT model, analyzed model components.
result Sudden drop in loss despite no changes in training procedure or hyper-parameters.
Transformers predict price movements from limit order books.
problem Predicting price movements from limit order books.
method Causal convolutional network with masked self-attention.
result Significantly outperforms existing architectures on FI-2010 dataset.
New method reduces diffusion model function evaluations for discrete data.
problem High computational burden in generating samples from masked diffusion models.
method Modified causal attention mask and speculative sampling mechanism for non-factorized predictions.
result Achieved ~2x reduction in required network forward passes.
Improved language models with talking-heads attention.
problem Language model perplexity and quality issues.
method Added linear projections in multi-head attention.
result Better perplexities and quality in language tasks.
Framework improves marine mammal monitoring in noisy underwater environments.
problem Underwater bioacoustic monitoring challenges due to overlapping calls and variable noise.
method Multi-step attention-guided framework with segmentation and mid-level fusion.
result Improved signal discrimination, reduced false positives, reliable representations.
Language models trained on chess board states outperform those on moves, even with causal masking.
problem Applying causal masking to spatial data for training unimodal language models.
method Trained bidirectional and causal self-attention models on both spatial (board-based) and sequential (move-based) chess data.
result Models trained on spatial board states achieve stronger playing strength than those trained on sequential data, even with causal masking.
Catastrophic forgetting occurs when a neural network loses the information learned in a previous task after training on subsequent tasks. This problem remains a hurdle for artificial intelligence systems with sequential learning capabilities. In this paper, we propose a task-based hard attention mechanism that preserve…
SPADE improves demand forecasting accuracy by 4.5% for post-promotion periods.
problem Overreacting to peak events in demand forecasting leads to biased forecasts.
method SPADE splits forecasting into two tasks: one for peak events and another for post-peak events, using masked convolution filters and a specialized Peak Attention module.
result Overall PPE improvement of 4.5%, 30% improvement for most affected forecasts after promotions and holidays, and 3.9% improvement in PE accuracy.
DOS improves language model generation by considering inter-token dependencies.
problem Lack of sequence-level information and inter-token dependencies in existing decoding strategies.
method Dependency-Oriented Sampler (DOS) that uses attention matrices to approximate inter-token dependencies.
result DOS consistently achieves superior performance on code generation and mathematical reasoning tasks.
I-BERT extends Transformer's self-attention to arbitrary input lengths.
problem Transformer models struggle with inductive generalization to unseen input lengths.
method Replaces positional encodings with a recurrent layer.
result I-BERT achieves state-of-the-art results on algorithmic tasks.
Recent variants improve knowledge distillation performance.
problem Improving the performance of knowledge distillation.
method Introducing additional components or changing the learning process.
result These variants have shown promising results.
Stochastic attention learns to retrieve and generate from memory without training.
problem Learning to retrieve and generate from memory efficiently.
method Langevin dynamics on modern Hopfield energy for stochastic attention.
result Stochastic attention can retrieve and generate from memory without training, matching gold standards.
Proposes a new approach to time series representation learning by embedding patches independently.
problem Capturing dependencies between time series patches is not optimal for representation learning.
method 1) Patch reconstruction task, 2) Patch-wise MLP, 3) Complementary contrastive learning.
result Improves time series forecasting and classification performance compared to state-of-the-art models.
Enhances time-series modeling by dropping patches, improving efficiency and adaptability.
problem Improving masked time-series modeling efficiency and adaptability.
method DropPatch method, which randomly drops sub-sequence level patches.
result DropPatch improves pre-training efficiency and adaptability in various learning scenarios.
CSHT predicts financial returns from news using a novel transformer model on a sphere.
problem Financial forecasting from news and sentiment.
method Granger-causal hypergraph structure, Riemannian geometry, causally masked Transformer attention.
result CSHT outperforms baselines in return prediction, regime classification, and asset ranking.
Acoustic scenes are rich and redundant in their content. In this work, we present a spatio-temporal attention pooling layer coupled with a convolutional recurrent neural network to learn from patterns that are discriminative while suppressing those that are irrelevant for acoustic scene classification. The convolutiona…
AaSP improves audio self-supervised learning by addressing aliasing issues.
problem Alias issues in audio spectrogram transformers.
method AaSP combines aliasing-aware patch representation, teacher-student masked modeling, cross-attention predictor, and contrastive regularization.
result AaSP learns more stable representations that integrate high-frequency cues.
This work explains the structural origins of attention sinks in LLMs.
problem Initial tokens disproportionately monopolize attention scores in LLMs.
method Traced to self-attention's value aggregation process and FFN layer activations.
result Attention sinks form due to variance discrepancy and dimension disparity.
PMI-Masking improves MLM pretraining by masking correlated spans efficiently.
problem Uniform token masking leads to inefficient and suboptimal performance in MLMs.
method PMI-Masking uses Pointwise Mutual Information to mask n-grams with high collocation.
result PMI-Masking reaches half the training time and improves performance.
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…
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.
We introduce a pretraining technique called Selfie, which stands for SELFie supervised Image Embedding. Selfie generalizes the concept of masked language modeling of BERT (Devlin et al., 2019) to continuous data, such as images, by making use of the Contrastive Predictive Coding loss (Oord et al., 2018). Given masked-o…
Paper proposes a method to encrypt faces while maintaining visual similarity.
problem Protecting personal data from unauthorized face recognition.
method Targeted identity-protection iterative method (TIP-IM) to generate adversarial identity masks.
result TIP-IM provides 95%+ protection success rate against face recognition models.
We present graph attention networks (GATs), novel neural network architectures that operate on graph-structured data, leveraging masked self-attentional layers to address the shortcomings of prior methods based on graph convolutions or their approximations. By stacking layers in which nodes are able to attend over thei…
Bidirectional attention is shown to be equivalent to a continuous bag of words model with mixture-of-experts.
problem Understanding the statistical underpinnings of bidirectional attention.
method Exploring bidirectional attention as a mixture-of-experts model and reparameterizing it.
result Bidirectional attention can be viewed as a continuous bag of words model with mixture-of-experts weights.
Paper proposes HGAT for detecting fake news on HIN.
problem Viral spread of fake news causing social harm.
method Hierarchical Graph Attention Network (HGAT) for node representation learning in HIN.
result HGAT outperforms text-based and network-based models.
ST-MTM models complex time series by decomposing and masking seasonal and trend components.
problem Forecasting complex time series with intricate temporal variations.
method Seasonal-Trend Decomposition with Masking and Contrastive Learning.
result ST-MTM achieves superior forecasting performance compared to existing methods.
LARA forecasts financial asset trends by refining noisy labels and extracting profitable samples.
problem Low signal-to-noise ratio and stochastic nature of financial data lead to poor predictions.
method LARA combines LA-Attention and RA-Labeling to refine and extract profitable samples.
result LARA significantly outperforms existing methods on Qlib platform.
Study examines asset pricing using various attention models, finding global self-attention and sliding window sparse attention models perform well.
problem Traditional asset pricing models miss temporal dependency and short memory issues.
method Investigates RNN attention models with various attention mechanisms for large-cap US stocks.
result Global self-attention and sliding window sparse attention models outperform in deriving returns and hedging risks, especially during the pandemic.
FMix improves model performance by distorting learned functions without distorting data distribution.
problem Improving model robustness and generalization performance in mixed sample data augmentation.
method Proposes FMix, an MSDA using random binary masks generated from Fourier space.
result FMix outperforms MixUp and CutMix, achieving state-of-the-art results on CIFAR-10.
Expands MLM by masking token positions, improving performance and convergence.
problem Improving language model performance and convergence.
method Masking token positions along with [MASK] tokens, using a fully connected classifier stage.
result Shows .3% improvement and 50% faster convergence for BERT Base with position masking.
SiamJEPA uses Siamese student encoders to improve JEPA-based representation learning.
problem Improving self-supervised representation learning in JEPA models.
method Proposes SiamJEPA with masked Siamese student encoders and EMA teacher network.
result Siamese student encoders improve representation separability and learning speed.
SMART training improves mask-predict translations.
problem Closing the performance gap between semi-autoregressive and autoregressive models.
method SMART training method for conditional masked language models.
result SMART-trained models produce higher-quality translations.
With widespread adoption of electronic health records, there is an increased emphasis for predictive models that can effectively deal with clinical time-series data. Powered by Recurrent Neural Network (RNN) architectures with Long Short-Term Memory (LSTM) units, deep neural networks have achieved state-of-the-art resu…
New analysis reveals masked self-supervised learning's effectiveness in extracting data structure.
problem Analyzing masked self-supervised learning in high-dimensional data.
method Developed precise high-dimensional analysis of masked modeling objectives.
result Identified phase transitions and structured regimes for masked self-supervised learning.
Develops a mean-field theory for multi-head self-attention under cross-entropy training.
problem Mean-field analysis of multi-head self-attention under cross-entropy training.
method Mean-field theory for a simplified single-layer causal multi-head self-attention model.
result Proves a static finite-head approximation bound for the optimal risk.