paper-with-me

Papers

Strassen Attention: Unlocking Compositional Abilities in Transformers Based on a New Lower Bound Method

2025-01-31 · Alexander Kozachinskiy, Felipe Urrutia, Hector Jimenez, Tomasz Steifer, Germán Pizarro, Matías Fuentes, Francisco Meza, Cristian Buc, Cristóbal Rojas

We propose a novel method to evaluate the theoretical limits of Transformers, allowing us to prove the first lower bounds against one-layer softmax Transformers with infinite precision. We establish those bounds for three tasks that require advanced reasoning. The first task, Match3 (Sanford et al., 2023), requires looking at all triples of positions. The second and third tasks address compositionality-based reasoning: one is composition of functions (Peng et al., 2024) and the other is composition of binary relations. We formally prove the inability of one-layer softmax Transformers to solve any of these tasks. In an attempt to overcome these limitations, we introduce Strassen attention and prove that with this mechanism a one-layer Transformer can in principle solve all these tasks. We also show that it enjoys sub-cubic running-time complexity, making it more scalable than similar previously proposed mechanisms, such as higher-order attention (Sanford et al., 2023). To complement our theoretical findings, we experimentally studied Strassen attention and compared it against standard (Vaswani et al, 2017), higher-order attention (Sanford et al., 2023) and triangular attention (Bergen et al. 2021). Our results help to disentangle all these attention mechanisms, highlighting their strengths and limitations. In particular, Strassen attention outperforms standard attention significantly on all the tasks. Altogether, understanding the theoretical limitations can guide research towards scalable attention mechanisms that improve the reasoning abilities of Transformers.

📄 PDF Abstract BibTeX arXiv:2501.19215

Code (0)

등록된 구현이 없습니다.

Tasks

All

Methods 이 논문이 사용한 방법론

Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Dense Connections Dense Connections, or Fully Connected Connections, are a type of layer in a deep neural network that use a linear operation where every input is connected to every output…
Residual Connection 설명 없음
Multi-Head Attention 설명 없음
Label Smoothing Label Smoothing is a regularization technique that introduces noise for the labels. This accounts for the fact that datasets may have mistakes in them, so maximizing the…
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
Position-Wise Feed-Forward Layer 설명 없음

Similar Papers 제목 키워드 기반

Compositional Capabilities of Autoregressive Transformers: A Study on Synthetic, Interpretable Tasks

2023-11-21 · Rahul Ramesh, Ekdeep Singh Lubana, Mikail Khona, Robert P. Dick 외

Transformers trained on huge text corpora exhibit a remarkable set of capabilities, e.g., performing basic arithmetic. Given the inherent compositional nature of language, one can expect the model to learn to compose the…

Unlocking Out-of-Distribution Generalization in Transformers via Recursive Latent Space Reasoning

2025-10-15 · Awni Altabaa, Siyu Chen, John Lafferty, Zhuoran Yang arxiv

Systematic, compositional generalization beyond the training distribution remains a core challenge in machine learning -- and a critical bottleneck for the emergent reasoning abilities of modern language models. This wor…

When Can Transformers Ground and Compose: Insights from Compositional Generalization Benchmarks

2022-10-23 · Ankur Sikarwar, Arkil Patel, Navin Goyal

Humans can reason compositionally whilst grounding language utterances to the real world. Recent benchmarks like ReaSCAN use navigation tasks grounded in a grid world to assess whether neural models exhibit similar capab…

Compositional Generalization (AVG)

Layer Specialization Underlying Compositional Reasoning in Transformers

2025-10-20 · Jing Liu arxiv

Transformers exhibit compositional reasoning on sequences not observed during training, a capability often attributed to in-context learning (ICL) and skill composition. We investigate this phenomenon using the Random Hi…

On the Computational Hardness of Transformers

2026-03-11 · Barna Saha, Yinzhan Xu, Christopher Ye, Hantao Yu arxiv

The transformer has revolutionized modern AI across language, vision, and beyond. It consists of $L$ layers, each running $H$ attention heads in parallel and feeding the combined output to the subsequent layer. In attent…