paper-with-me

홈 › Papers

BERTology Meets Biology: Interpreting Attention in Protein Language Models

2020-06-26 · ICLR 2021 1 · Jesse Vig, Ali Madani, Lav R. Varshney, Caiming Xiong, Richard Socher, Nazneen Fatema Rajani

Transformer architectures have proven to learn useful representations for protein classification and generation tasks. However, these representations present challenges in interpretability. In this work, we demonstrate a set of methods for analyzing protein Transformer models through the lens of attention. We show that attention: (1) captures the folding structure of proteins, connecting amino acids that are far apart in the underlying sequence, but spatially close in the three-dimensional structure, (2) targets binding sites, a key functional component of proteins, and (3) focuses on progressively more complex biophysical properties with increasing layer depth. We find this behavior to be consistent across three Transformer architectures (BERT, ALBERT, XLNet) and two distinct protein datasets. We also present a three-dimensional visualization of the interaction between attention and protein structure. Code for visualization and analysis is available at https://github.com/salesforce/provis.

📄 PDF Abstract BibTeX arXiv:2006.15222

Code (2)

salesforce/provis 공식 구현 pytorch
tadorfer/NLProt pytorch

Methods 이 논문이 사용한 방법론

Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Absolute Position Encodings Absolute Position Encodings are a type of position embeddings for [Transformer-based models] where positional encodings are…
Position-Wise Feed-Forward Layer 설명 없음
WordPiece 설명 없음
Refunds@Expedia|||How do I get a full refund from Expedia? “How do I get a full refund from Expedia? How do I get a full refund from Expedia? – Call ☎️ +1-(888) 829 (0881) or +1-805-330-4056 or +1-805-330-4056 for Quick Help &…
LAMB LAMB is a a layerwise adaptive large batch optimization technique. It provides a strategy for adapting the learning rate in large batch settings. LAMB uses…
ALBERT 설명 없음
Residual Connection 설명 없음

Similar Papers 제목 키워드 기반

Flow Matching Meets Biology and Life Science: A Survey

2025-07-23 · Zihao Li, Zhichen Zeng, Xiao Lin, Feihao Fang 외 arxiv

Over the past decade, advances in generative modeling, such as generative adversarial networks, masked autoencoders, and diffusion models, have significantly transformed biological research and discovery, enabling breakt…

Drug Discovery

Induction Meets Biology: Mechanisms of Repeat Detection in Protein Language Models

2026-02-26 · Gal Pomerants, Yaniv Nikankin, Anja Reusch, Tomer Tsaban 외 arxiv

Protein sequences are abundant in repeating segments, both as exact copies and as approximate segments with mutations. These repeats are important for protein structure and function, motivating decades of algorithmic wor…

Incorporating network based protein complex discovery into automated model construction

2020-09-29 · Paul Scherer, Maja Trȩbacz, Nikola Simidjievski, Zohreh Shams 외

We propose a method for gene expression based analysis of cancer phenotypes incorporating network biology knowledge through unsupervised construction of computational graphs. The structural construction of the computatio…

Clustering

Recent advances in interpretable machine learning using structure-based protein representations

2024-09-26 · Luiz Felipe Vecchietti, Minji Lee, Begench Hangeldiyev, Hyunkyu Jung 외

Recent advancements in machine learning (ML) are transforming the field of structural biology. For example, AlphaFold, a groundbreaking neural network for protein structure prediction, has been widely adopted by research…

Interpretable Machine LearningProtein DesignProtein Structure Prediction

GPCR-BERT: Interpreting Sequential Design of G Protein Coupled Receptors Using Protein Language Models

2023-10-30 · Seongwon Kim, Parisa Mollaei, Akshay Antony, Rishikesh Magar 외

With the rise of Transformers and Large Language Models (LLMs) in Chemistry and Biology, new avenues for the design and understanding of therapeutics have opened up to the scientific community. Protein sequences can be m…