Related papers: LUNA: Linear Universal Neural Attention with Gener…

We find arithmetic ability resides within a limited number of attention heads, with each head specializing in distinct operations. To delve into the reason, we introduce the Comparative Neuron Analysis (CNA) method, which identifies an…

Feature selection is the problem of selecting a subset of features for a machine learning model that maximizes model quality subject to a budget constraint. For neural networks, prior methods, including those based on $\ell_1$…

We introduce Performers, Transformer architectures which can estimate regular (softmax) full-rank-attention Transformers with provable accuracy, but using only linear (as opposed to quadratic) space and time complexity, without relying on…

Large language models (LLMs) enable long-context tasks but face efficiency challenges due to the growing key-value (KV) cache. We propose LeanK, a learning-based method that prunes unimportant key (K) cache channels by leveraging static…

Many efficient $\textit{approximate}$ self-attention techniques have become prevalent since the inception of the transformer architecture. Two popular classes of these techniques are low-rank and kernel methods. Each of these methods has…

Autoregressive (AR) video diffusion is a powerful paradigm for streaming and interactive video generation. However, its reliance on softmax self-attention leads to quadratic compute complexity in sequence length and memory usage due to…

Linear attentions have shown potential for improving Transformer efficiency, reducing attention's quadratic complexity to linear in sequence length. This holds exciting promise for (1) training linear Transformers from scratch, (2)…

Training causal transformers at extreme sequence lengths is bottlenecked by the quadratic time and memory of scaled dot-product attention (SDPA). In this work, we propose Lighthouse Attention, a training-only symmetrical selection-based…

Transformers have become the de facto standard for a wide range of tasks, from image classification to physics simulations. Despite their impressive performance, the quadratic complexity of standard Transformers in both memory and time with…

Previous research has explored the computational expressivity of Transformer models in simulating Boolean circuits or Turing machines. However, the learnability of these simulators from observational data has remained an open question. Our…

Transformers with linear attention (i.e., linear transformers) and state-space models have recently been suggested as a viable linear-time alternative to transformers with softmax attention. However, these models still underperform…

Large Language Model (LLM) deployment is increasingly shifting to cost-efficient accelerators like Google's Tensor Processing Units (TPUs), prioritizing both performance and total cost of ownership (TCO). However, existing LLM inference…

Convolutional neural networks (CNNs) have demonstrated superior performance in super-resolution (SR). However, most CNN-based SR methods neglect the different importance among feature channels or fail to take full advantage of the…

We present TransNormerLLM, the first linear attention-based Large Language Model (LLM) that outperforms conventional softmax attention-based models in terms of both accuracy and efficiency. TransNormerLLM evolves from the previous linear…

Most existing structured pruning methods for Large Language Models (LLMs) require substantial computational and data resources for retraining to reestablish the corrupted correlations, making them prohibitively expensive. To address this,…

Large Language Models (LLMs) have demonstrated remarkable capabilities across a wide range of natural language processing tasks. These capabilities stem primarily from the self-attention mechanism, which enables modeling of long-range…

Sequential recommendation models, particularly those based on attention, achieve strong accuracy but incur quadratic complexity, making long user histories prohibitively expensive. Sub-quadratic operators such as Hyena provide efficient…

We prove that with linear transformations, both (i) two-layer self-attention and (ii) one-layer self-attention followed by a softmax function are universal approximators for continuous sequence-to-sequence functions on compact domains. Our…

State-of-the-art sequential recommendation relies heavily on self-attention-based recommender models. Yet such models are computationally expensive and often too slow for real-time recommendation. Furthermore, the self-attention operation…

Many advanced Large Language Model (LLM) applications require long-context processing, but the self-attention module becomes a bottleneck during the prefilling stage of inference due to its quadratic time complexity with respect to sequence…