New insights into attention mechanisms reveal dramatic trade-offs between rank and heads.
problem Dramatic trade-offs between rank and number of heads in attention mechanisms.
method Presented a simple target function and proved theoretical limits.
result Full-rank attention is necessary for long contexts, while low-rank is sufficient for short ones.
New approach reduces model size for Transformer architectures.
problem Large embedding dimensions limit model applicability.
method Identified low-rank bottleneck in multi-head attention.
result Reducing head size to sequence length improves model performance.
Transformers without skip connections collapse token representations to a single direction.
problem Rapid convergence of token representations to a single direction in self-attention-only Transformers.
method Analysis of layer normalization, residual connections, and multi-head attention mechanisms.
result Residual connections prevent rank collapse in real Transformers, while MLPs generate new feature directions.
We show the rank (i.e. minimal size of a generating set) of lattices cannot grow faster than the volume.
Choppy optimizes ranked list truncation using Transformer architecture.
problem Optimal truncation of ranked search results to balance relevance and user cost.
method Assumption-free Transformer model optimizing user-defined IR metrics.
result Choppy improves upon recent state-of-the-art methods.
We observe that stable integral simplicial volume of closed manifolds gives an upper bound for the rank gradient of the corresponding fundamental groups.
Proposes a new neural head for asymmetric representation learning.
problem Asymmetric representation learning in directed relations.
method Role-aware neural convex divergence head.
result Role-aware projections improve directional accuracy over plain ICNN-Bregman heads.
We study the growth of the rank of subgroups of finite index in residually finite groups, by relating it to the notion of cost. As a by-product, we show that the `Rank vs. Heegaard genus' conjecture on hyperbolic 3-manifolds is incompatible with the `Fixed Price problem' in topological dynamics.
An important research direction in machine learning has centered around developing meta-learning algorithms to tackle few-shot learning. An especially successful algorithm has been Model Agnostic Meta-Learning (MAML), a method that consists of two optimization loops, with the outer loop finding a meta-initialization, f…
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.
Global models outperform local models in forecasting intermittent time series.
problem Forecasting intermittent time series with zeros in supply chains.
method Comparison of state-of-the-art probabilistic local and global models on five datasets.
result TiDE, a simple neural network architecture, achieves the best accuracy among global models.
GWHD dataset offers 4,700 high-res images of wheat heads.
problem Challenges in wheat head detection from high-resolution imagery.
method Large, diverse dataset with detailed metadata.
result Benchmark for wheat head detection methods.
Multi-head attention outperforms single-head in in-context linear regression tasks.
problem Comparing performance of transformer with single-/multi-head attention in in-context learning.
method Theoretical analysis of performance of transformers with different attention mechanisms in linear regression tasks.
result Multi-head attention with a substantial embedding dimension outperforms single-head attention in in-context linear regression tasks.
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.
New theory shows how multi-head attention reduces variance and decorrelates outputs.
problem Understanding and optimizing multi-head attention in neural networks.
method Developed a statistical theory linking multi-head attention to ensemble Nadaraya-Watson estimators.
result MHA variance reduction depends on head decorrelation, not just head count.
Geometric theory of projection heads in self-supervised learning.
problem Dimensional collapse and information invariance trade-off in projection heads.
method Geometric modeling of projection heads as Riemannian metrics, analyzing Hessian eigenvalues, and tracking optimization geometry.
result Smooth nonlinear heads induce negative curvature, preventing collapse; linear and ReLU heads cannot.
Gated attention improves performance by using a hierarchical mixture of experts.
problem Improving performance of self-attention mechanisms in Transformers.
method Rigorously show that gated attention can be modeled as a hierarchical mixture of experts, providing a theoretical justification for its benefits.
result Gated attention is more sample-efficient than multi-head self-attention, requiring fewer data points to achieve the same estimation error.
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.
New approach improves multi-head attention by making heads less similar.
problem Multi-head attention can lead to similar features, reducing model expressiveness.
method Proposes a non-parametric approach using Bayesian techniques to make heads repel each other.
result Improves feature diversity, leading to better representations and performance.
Paper extends multivariate rank tests for robust subspace detection.
problem Testing distributional similarity in multivariate data.
method Soft and subspace robust multivariate rank tests based on entropy regularized optimal transport.
result Trade-off between detection power and false alarm rate via projections.
TSFMs show redundant components in layers, affecting their performance.
problem Redundant components in TSFMs layers impact their predictive accuracy.
method Mechanistic interpretability tools, ablations, logit attribution, theoretical framework.
result TSFMs are robust to ablations of entire layers and specific heads.
Deep forecasting models show output heads significantly improve performance on fat-tailed financial returns.
problem Improving deep learning models for forecasting fat-tailed financial returns.
method Comparison of backbone architectures and output heads (point, Gaussian, Gaussian mixture) on S&P 500 monthly log-returns.
result Switching from point to Gaussian heads improves CRPS by about 1.3 percent, and from Gaussian to mixture adds another 2.4 percent.
LEMs extend transformer-based architectures for complex execution problems.
problem Handling flexible time boundaries and multiple execution constraints in deep learning.
method Decouples market information processing from execution allocation decisions using TKANs, VSNs, and multi-head attention mechanisms.
result LEMs achieve superior execution performance compared to traditional benchmarks.
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.
Multi-head attention mechanism is capable of learning various representations from sequential data while paying attention to different subsequences, e.g., word-pieces or syllables in a spoken word. From the subsequences, it retrieves richer information than a single-head attention which only summarizes the whole sequen…
A graph neural network improves multivariate post-processing of ensemble forecasts.
problem Systematic biases in ensemble forecasts and loss of dependencies across forecast dimensions.
method A composite-Loss Graph Neural Network (dualGNN) trained with a composite loss function combining ES and VS.
result The dualGNN outperforms traditional methods in multivariate verification metrics and captures spatial relationships.
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…
One-layer transformers can't solve induction heads task efficiently.
problem Solving the induction heads task efficiently with one-layer transformers.
method Communication complexity argument showing exponential size requirement.
result No one-layer transformer can solve the induction heads task efficiently.
Survey of Turk's head knots and links properties.
problem Characterize Turk's head knots and links.
method Discussion of various results in the mathematical literature.
result Turk's head links are alternating, fibered, hyperbolic, invertible, non-split, periodic, and prime.
This work proposes a collaborative multi-head attention layer to reduce model size without sacrificing accuracy.
problem Over-parameterization in transformer models trained with large datasets.
method Proposes a collaborative multi-head attention layer that shares key/query projections.
result Reduction in model size by 4 for same accuracy and speed.
This paper reviews methods for discovering patient subgroups from EHR data.
problem Discovering subgroups of patients and co-occurring medical conditions from EHR data.
method Low-rank data approximation methods like matrix and tensor decompositions.
result These methods provide transparent and interpretable insights into patient phenotypes.
Study compares and contrasts various ML explanation methods, highlighting their disagreements and similarities.
problem Understanding and quantifying the differences between various machine learning explanation methods.
method Synthesized and visualized various explanation methods for global and local aspects of ML models.
result There is substantial agreement on the top features but less on specific rankings, and tree interpreter is comparable to SHAP for feature effects.
Multi-headed ensembles boost model performance with faster training.
problem Limited computational resources hinder ensemble search performance.
method Extend NES to multi-headed ensembles, leveraging end-to-end training and one-shot NAS methods.
result Multi-headed ensemble search finds robust ensembles 3 times faster with comparable performance.
Study on multi-head softmax attention dynamics for in-context learning.
problem Understanding and optimizing multi-head softmax attention models for multi-task linear regression.
method Gradient flow analysis and spectral mapping technique.
result Gradient flow converges to optimal multi-head softmax attention model, with task allocation emerging during training.
Transformer-MGK replaces redundant heads with Gaussian key mixtures, improving efficiency and performance.
problem Redundant attention heads in transformers degrade performance and efficiency.
method Transformer-MGK replaces redundant heads with a mixture of Gaussian keys.
result Transformer-MGK accelerates training and inference, reduces parameters and FLOPs, and achieves comparable or better accuracy.
Minimalistic model captures head direction system properties.
problem Representing head direction system in a high-dimensional space.
method A minimalistic representation model of the rotation group U(1), including fully connected and convolutional versions.
result Emergence of Gaussian-like tuning profiles and 2D circle geometry in both model versions.
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.
FUSE neural centrality framework improves data point measurement in high dimensions.
problem Measuring centrality in high-dimensional data is expensive and unstable.
method Combines global and local heads trained on arbitrary representations.
result Reveals meaningful classical ordering and competitive performance.
Investigates the benefits of multi-head attention in Transformers, deriving convergence and generalization guarantees.
problem Underexplored dynamics of multi-head attention in Transformer training and generalization.
method Derives convergence and generalization guarantees for gradient-descent training of a multi-head self-attention model.
result Establishes conditions for initialization that ensure multi-head attention's realizability.
We study the head and tail of the colored Jones polynomial while focusing mainly on alternating links. Various ways to compute the colored Jones polynomial for a given link give rise to combinatorial identities for those power series. We further show that the head and tail functions only depend on the reduced checkerbo…
Transformer learns to search through reinforcement learning, mimicking DFS.
problem Understanding how transformers learn search capabilities in RL.
method Two-head transformer, depth-wise curriculum, discounted returns.
result Transformer policy generalizes depth and prioritizes high-probability branches.
New deep learning model generates accurate personalized human head models for electromagnetic dosimetry.
problem Challenges in generating accurate human head models for personalized electromagnetic dosimetry.
method Proposed ForkNet architecture for segmentation of whole human head structures using deep learning.
result Generated head models exhibit strong matching with manual segmentation results.
Investigates optimal parameter allocation in Transformers for efficiency and expressivity.
problem Balancing expressivity and efficiency in Transformer model parameters.
method Mathematical analysis and theoretical characterization of attention heads and head dimensions.
result Later layers can operate more efficiently with reduced parameters due to saturation of softmax activations.
We demystify attention patterns in multi-head softmax models for linear data.
problem Understanding the training dynamics and emergent patterns in multi-head softmax attention models.
method Extensive empirical experiments and rigorous theoretical analysis.
result Multi-head softmax attention models approximate a debiased gradient descent predictor, outperforming single-head attention and achieving near-Bayesian optimality.
Improved robot navigation using multi-head attention for natural language instructions.
problem Improving robot navigation in unfamiliar environments.
method Proposes a multi-head attention mechanism blending layer in a neural network model.
result Significant performance gains in translating instructions for unseen environments.
Study shows the number of attention heads affects transformer performance.
problem Understanding how the number of attention heads impacts transformer performance.
method Introduced a generalized D-retrieval task, established upper and lower bounds on parameter complexity, and validated with experiments. result Transformers with many heads can efficiently approximate functions, while few heads require a large number of parameters.
We show that the head and tail functions of the colored Jones polynomial of adequate links are the product of head and tail functions of the colored Jones polynomial of alternating links that can be read-off an adequate diagram of the link. We apply this to strengthen a theorem of Kalfagianni, Futer and Purcell on the …
New method removes contrastive loss by adding a prediction head, revealing learning mechanisms.
problem Understanding why neural networks learn competitive representations despite trivial optima.
method Empirical and theoretical analysis of a trainable, identity-initialized prediction head.
result The trainable prediction head enables learning all features, preventing dimensional collapse.