Challenge: Pre-trained transformers suffer from a computationally expensive self-attention mechanism that interacts with all tokens, including those unfavorable to classification performance.
Approach: They propose to integrate token pruning and token combining strategies to improve model performance and reduce computational demands.
Outcome: Experiments with various datasets show that the proposed model performs better than baseline models, with the best improvement over the existing model.

Similar Papers

Token and Head Adaptive Transformers for Efficient Natural Language Processing (2022.coling-1)

Copied to clipboard

Challenge: Pre-trained language models like BERT have shown significant accuracy improvements on various tasks, but their computational cost and memory footprint are prohibitive.
Approach: They propose to extend Length Adaptive Transformer to extend the model to a token and head pruning scheme to optimize pruning efficiency.
Outcome: The proposed model can compress and accelerate BERT-based models by fine-tuning and a token and head pruning scheme.
DoT: An efficient Double Transformer for NLP tasks with tables (2021.findings-acl)

Copied to clipboard

Challenge: Recent studies have shown that transformer-based approaches to NLP tasks are slow and require computational and memory costs.
Approach: They propose a transformer-based model that decomposes a problem into two sub-tasks and a pruning transformer that takes as input the pruning scores.
Outcome: The proposed model improves training and inference time by at least 50% for a small drop in accuracy and also enables the model to maintain similar accuracy as slower baseline models.
Rethinking Network Pruning – under the Pre-train and Fine-tune Paradigm (2021.naacl-main)

Copied to clipboard

Challenge: Existing pruning results on benchmark transformers, such as BERT, are not as remarkable as those of convolutional neural networks.
Approach: They propose to apply a knowledge-aware pruning process to transformer-based pre-trained language models to reduce model size and model weight.
Outcome: The proposed pruning method outperforms the leading competitors with a 20-times weight/FLOPs compression and neglectable loss in prediction accuracy.
Compressing Large-Scale Transformer-Based Models: A Case Study on BERT (2021.tacl-1)

Copied to clipboard

Challenge: Popular pre-trained Transformers have improved performance for various NLP tasks by sizable margins, but are too resource-hungry and computation-intensive to suit low-capacity devices or applications with strict latency requirements.
Approach: They present a literature review of the compression of Transformers, focusing on the popular BERT model, which has attracted considerable research attention.
Outcome: The proposed models improve Sentiment analysis, paraphrase detection, machine reading comprehension, question answering, text summarization, and other tasks by sizable margins.
Symmetric Dot-Product Attention for Efficient Training of BERT Language Models (2024.findings-acl)

Copied to clipboard

Challenge: Transformer-based models are stretched to enormous sizes, requiring increasingly larger training datasets and unsustainable amount of compute resources.
Approach: They propose an alternative compatibility function for the Transformer-based attention mechanism that exploits an overlap in the learned representation of the traditional scaled dot-product attention mechanism.
Outcome: The proposed model achieves 79.36 on the GLUE benchmark against 78.74 for the traditional implementation and reduces the number of trainable parameters by 6%.
SmartTrim: Adaptive Tokens and Attention Pruning for Efficient Vision-Language Models (2024.lrec-main)

Copied to clipboard

Challenge: Experimental results show that SmartTrim accelerates the original model by 2-3 times with minimal performance degradation.
Approach: They propose an adaptive acceleration framework which prunes redundant token representations and attention heads within each layer of the original model.
Outcome: The proposed framework accelerates the original model by 2-3 times with minimal performance degradation across vision-language tasks.
Context Compression for Auto-regressive Transformers with Sentinel Tokens (2023.emnlp-main)

Copied to clipboard

Challenge: Existing Transformer-based LLMs have limited performance due to complexity of attention module . key-value cache is the major memory footprint and inference latency problem .
Approach: They propose a plug-and-play approach that incrementally compresses token activation into compact ones . they also profile the benefit of context compression on improving the system throughout .
Outcome: The proposed approach reduces memory footprint and inference latency by compressing tokens into compact ones.
Integral Transformer: Denoising Attention, Not Too Much Not Too Little (2025.emnlp-main)

Copied to clipboard

Challenge: Existing methods to reduce attention noise by integrating signals from logit distributions are prone to attention noise.
Approach: They propose a self-attention mechanism that integrates signals from the logit distribution to denoise attention.
Outcome: The proposed model outperforms vanilla, Cog, and Differential attention variants on knowledge and reasoning benchmarks.
Fine- and Coarse-Granularity Hybrid Self-Attention for Efficient BERT (2022.acl-long)

Copied to clipboard

Challenge: Transformer-based pre-trained models achieve state-of-the-art results, but they can be prohibitively costly.
Approach: They propose a fine- and coarse-granularity hybrid self-attention that shortens the computational sequence length in self- attention by progressively shortening the computational time.
Outcome: The proposed model reduces computation cost by shortening the computational sequence length in self-attention.
Fine-Tuning Pre-trained Transformers into Decaying Fast Weights (2022.emnlp-main)

Copied to clipboard

Challenge: Autoregressive Transformers incur O(T) complexity during per-token generation due to the self-attention mechanism.
Approach: They propose a kernel-based method to approximate causal self-attention by replacing it with recurrent formulations with various update rules and feature maps to achieve O(1) time and memory complexity.
Outcome: The proposed method outperforms prior methods and retains 99% of attention’s performance on WikiText-103 against more complex attention substitutes.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations