Overview

Gated attention studies how learned controls can regulate information flow in Transformers and related models. Common approaches use multiplicative gates on attention outputs, values, heads, or tokens, while gated linear and recurrent methods use input-dependent updates or decay to handle long sequences efficiently. These methods aim to improve selectivity, robustness, and computational cost, with applications spanning language modeling, long-context retrieval, vision, graphs, and multimodal tasks. Results vary by task and scale, and an open challenge is separating gains from gating itself from model adaptation, while checking whether soft learned gates deliver reliable benefits when converted to hard pruning or sparse inference.

Latest papers

37 papers4 papers in the last four weeks, down 56% on the four weeks before. 0.0% of all new papers.

Apr 13, 2026

Learning to Adapt: In-Context Learning Beyond Stationarity

Transformer models have become foundational across a wide range of scientific and engineering domains due to their strong empirical performance. A key capability underlying their success is in-context learning (ICL): when presented with a short prompt from an unseen task, transformers can perform per-token and next-token predictions without any parameter updates. Recent theoretical efforts have begun to uncover the mechanisms behind this phenomenon, particularly in supervised regression settings. However, these analyses predominantly assume stationary task distributions, which overlook a broad class of real-world scenarios where the target function varies over time. In this work, we bridge this gap by providing a theoretical analysis of ICL under non-stationary regression problems. We study how the gated linear attention (GLA) mechanism adapts to evolving input-output relationships and rigorously characterize its advantages over standard linear attention in this dynamic setting. To model non-stationarity, we adopt a first-order autoregressive process and show that GLA achieves lower training and testing errors by adaptively modulating the influence of past inputs -- effectively implementing a learnable recency bias. Our theoretical findings are further supported by empirical results, which validate the benefits of gating mechanisms in non-stationary ICL tasks.
Apr 7, 2026

Gated-SwinRMT: Unifying Swin Windowed Attention with Retentive Manhattan Decay via Input-Dependent Gating

We introduce Gated-SwinRMT, a family of hybrid vision transformers that combine the shifted-window attention of the Swin Transformer with the Manhattan-distance spatial decay of Retentive Networks (RMT), augmented by input-dependent gating. Self-attention is decomposed into consecutive width-wise and height-wise retention passes within each shifted window, where per-head exponential decay masks provide a two-dimensional locality prior without learned positional biases. Two variants are proposed.Gated-SwinRMT-SWAT substitutes softmax with sigmoid activation, implements balanced ALiBi slopes with multiplicative post-activation spatial decay, and gates the value projection via SwiGLU; the Normalized output implicitly suppresses uninformative attention scores. Gated-SwinRMT-Retention retains softmax-normalized retention with an additive log-space decay bias and incorporates an explicit G1 sigmoid gate -- projected from the block input and applied after local context enhancement (LCE) but prior to the output projection~WOW_O -- to alleviate the low-rank WV ⁣⋅ ⁣WOW_V \!\cdot\! W_O bottleneck and enable input-dependent suppression of attended outputs. We assess both variants on Mini-ImageNet (224×224224{\times}224, 100 classes) and CIFAR-10 (32×3232{\times}32, 10 classes) under identical training protocols, utilizing a single GPU due to resource limitations. At ≈77{\approx}77--7979,M parameters, Gated-SwinRMT-SWAT achieves 80.22%80.22\% and Gated-SwinRMT-Retention 78.20%78.20\% top-1 test accuracy on Mini-ImageNet, compared with 73.74%73.74\% for the RMT baseline. On CIFAR-10 -- where small feature maps cause the adaptive windowing mechanism to collapse attention to global scope -- the accuracy advantage compresses from +6.48+6.48,pp to +0.56+0.56,pp.
Mar 7, 2026

The Model Knows Which Tokens Matter: Automatic Token Selection via Noise Gating

Visual tokens dominate inference cost in vision-language models (VLMs), yet many carry redundant information. Existing pruning methods alleviate this but typically rely on attention magnitude or similarity scores. We reformulate visual token pruning as capacity constrained communication: given a fixed budget K, the model must allocate limited bandwidth to maximally preserve visual information. We propose AutoSelect, which attaches a lightweight Scorer and Denoiser to a frozen VLM and trains with only the standard next token prediction loss, without auxiliary objectives or extra annotations. During training, a variance preserving noise gate modulates each token's information flow according to its predicted importance so that gradients propagate through all tokens; a diagonal attention Denoiser then recovers the perturbed representations. At inference, only the Scorer and a hard top-K selection remain, adding negligible latency. On ten VLM benchmarks, AutoSelect retains 96.5% of full model accuracy while accelerating LLM prefill by 2.85x with only 0.69 ms overhead, and transfers to different VLM backbones without architecture-specific tuning. Code is available at https://github.com/MedHK23/AutoSelect.
Mar 3, 2026

Gated Differential Linear Attention: A Linear-Time Decoder for High-Fidelity Medical Segmentation

Medical image segmentation requires models that preserve fine anatomical boundaries while remaining practical for clinical deployment. Transformers capture long-range dependencies but incur quadratic attention cost, whereas CNNs are efficient but less effective at global reasoning. Linear attention offers O(N)\mathcal{O}(N) scaling, but often produces diffuse feature aggregation that weakens boundary-sensitive prediction. We introduce a gated differential linear-attention mixer for medical image segmentation. Its global path, Gated Differential Linear Attention (GDLA), performs differential subtraction between two kernelized attention branches over complementary query/key subspaces to suppress redundant responses, and employs a data-dependent gate for token refinement. A parallel local token-mixing branch with depthwise convolution strengthens neighborhood interactions for better refinement, and the two branches are fused while preserving O(N)\mathcal{O}(N) complexity. When instantiated in a pretrained Pyramid Vision Transformer (PVT)-based encoder--decoder model, \name achieves state-of-the-art results on the evaluated 2D medical segmentation benchmarks spanning CT, MRI, ultrasound, and dermoscopy, with a favorable accuracy--efficiency trade-off over closely related baselines. The code is publicly available at https://github.com/xmindflow/gdla.
Feb 11, 2026

Routing Absorption in Sparse Attention: Why Random Gates Are Hard to Beat

Learned gates can approximate sparse attention patterns on frozen transformers, yet provide limited benefit over random gates when trained jointly with the model. We investigate this difference in a controlled 31M-parameter transformer and attribute it to routing absorption: model representations co-adapt to the imposed mask, reducing the incremental benefit of learned routing. Four experiments characterize the phenomenon. Differentiable soft gating yields perplexities of 48.73 plus or minus 0.60 with learned gates and 49.83 plus or minus 0.04 with frozen random gates over three seeds. Hard top-k masking provides no gradient path to the gate scores in the tested implementation. Gates distilled onto co-adapted and dense-trained Q/K/V both achieve high F1 against oracle masks, but hard-mask deployment yields perplexities of 601.6 and 48.6, respectively. Stochastic mask training also leaves a substantial deployment penalty: dense evaluation yields 78.2 perplexity, compared with 37.3 for the dense baseline. Experiments on Qwen3-1.7B show that increasing the number of trainable attention layers reduces the gap between learned and random gates. We relate these results to co-adaptation in Mixture-of-Experts and propose parameter asymmetry between the gate and the model as a contributing mechanism. For the tested per-query, token-level gates, freezing the model provides stable routing targets and enables effective post-hoc sparsification. The results motivate random-routing controls and separate evaluation of routing quality and model adaptation in sparse attention methods.
Feb 1, 2026

A Statistical Theory of Gated Attention through the Lens of Hierarchical Mixture of Experts

Self-attention has greatly contributed to the success of the widely used Transformer architecture by enabling learning from data with long-range dependencies. In an effort to improve performance, a gated attention model that leverages a gating mechanism within the multi-head self-attention has recently been proposed as a promising alternative. Gated attention has been empirically demonstrated to increase the expressiveness of low-rank mapping in standard attention and even to eliminate the attention sink phenomenon. Despite its efficacy, a clear theoretical understanding of gated attention's benefits remains lacking in the literature. To close this gap, we rigorously show that each entry in a gated attention matrix or a multi-head self-attention matrix can be written as a hierarchical mixture of experts. By recasting learning as an expert estimation problem, we demonstrate that gated attention is more sample-efficient than multi-head self-attention. In particular, while the former needs only a polynomial number of data points to recover the model parameters, the latter requires exponentially many data points to achieve the same estimation error. Furthermore, our analysis also provides a theoretical justification for why gated attention is more effective at mitigating the attention sink phenomenon and the low-rank bottleneck when a gate is placed at the output of the scaled dot product attention or the value map rather than at other positions in the multi-head self-attention architecture.
Aug 24, 2025

GateTS: Versatile and Efficient Forecasting via Attention-Inspired routed Mixture-of-Experts

Accurate univariate forecasting remains a pressing need in real-world systems, such as energy markets, hydrology, retail demand, and IoT monitoring, where signals are often intermittent and horizons span both short- and long-term. While transformers and Mixture-of-Experts (MoE) architectures are increasingly favored for time-series forecasting, a key gap persists: MoE models typically require complicated training with both the main forecasting loss and auxiliary load-balancing losses, along with careful routing/temperature tuning, which hinders practical adoption. In this paper, we propose a model architecture that simplifies the training process for univariate time series forecasting and effectively addresses both long- and short-term horizons, including intermittent patterns. Our approach combines sparse MoE computation with a novel attention-inspired gating mechanism that replaces the traditional one-layer softmax router. Through extensive empirical evaluation, we demonstrate that our gating design naturally promotes balanced expert utilization and achieves superior predictive accuracy without requiring the auxiliary load-balancing losses typically used in classical MoE implementations. The model achieves better performance while utilizing only a fraction of the parameters required by state-of-the-art transformer models, such as PatchTST. Furthermore, experiments across diverse datasets confirm that our MoE architecture with the proposed gating mechanism is more computationally efficient than LSTM for both long- and short-term forecasting, enabling cost-effective inference. These results highlight the potential of our approach for practical time-series forecasting applications where both accuracy and computational efficiency are critical.