Length Extrapolation of Transformers: A Survey from the Perspective of Positional Encoding (2024.findings-emnlp)
Copied to clipboard
Liang Zhao, Xiachong Feng, Xiaocheng Feng, Weihong Zhong, Dongliang Xu, Qing Yang, Hongtao Liu, Bing Qin, Ting Liu
| Challenge: | Existing methods to enhance length extrapolation of large language models have been developed, but a systematic survey is lacking. |
| Approach: | They propose to examine the effects of positional encoding on length extrapolation. |
| Outcome: | The proposed methods improve the extrapolation of large language models, but they are still lacking a systematic survey. |
Similar Papers
Randomized Positional Encodings Boost Length Generalization of Transformers (2023.acl-short)
Copied to clipboard
Anian Ruoss, Grégoire Delétang, Tim Genewein, Jordi Grau-Moya, Róbert Csordás, Mehdi Bennani, Shane Legg, Joel Veness
| Challenge: | Moreover, simply training on longer sequences is inefficient due to the quadratic computation complexity of the global attention mechanism. |
| Approach: | They propose a randomized positional encoding scheme that randomly selects an ordered subset to fit the sequence’s length. |
| Outcome: | The proposed method allows Transformers to generalize to sequences of unseen length (increasing test accuracy by 12.0% on average). |
Position Encoding with Random Float Sampling Enhances Length Generalization of Transformers (2026.findings-eacl)
Copied to clipboard
| Challenge: | Length generalization is the ability of language models to maintain performance on inputs longer than those seen during pretraining. |
| Approach: | They propose a position encoding strategy that uses random float sampling to generalize to unseen lengths. |
| Outcome: | The proposed strategy can generalize to lengths unseen during training and in benchmarks. |
A Length-Extrapolatable Transformer (2023.acl-long)
Copied to clipboard
Yutao Sun, Li Dong, Barun Patra, Shuming Ma, Shaohan Huang, Alon Benhaim, Vishrav Chaudhary, Xia Song, Furu Wei
| Challenge: | Existing Transformers can only deal with the in-distribution size of inputs. |
| Approach: | They propose a relative position embedding to explicitly maximize attention resolution . they also use blockwise causal attention during inference for better resolution a . |
| Outcome: | The proposed model achieves strong performance in interpolation and extrapolation settings. |
Sequence Length is a Domain: Length-based Overfitting in Transformer Models (2021.emnlp-main)
Copied to clipboard
| Challenge: | Current Transformer-based sequence-to-sequence architectures can suffer from overfitting during training. |
| Approach: | They propose to use Transformer-based sequence-to-sequence architectures to overcome overfitting problems when generating very long sequences. |
| Outcome: | The proposed model performs worse on very long sequences than previous approaches on string editing and translation tasks when faced with sequences of length diverging from the length distribution in training data. |
Shortformer: Better Language Modeling using Shorter Inputs (2021.acl-long)
Copied to clipboard
| Challenge: | Existing methods require computationally expensive relative position embeddings. |
| Approach: | They propose two methods that decrease input length to improve perplexity and perplexability. |
| Outcome: | The proposed methods speed up training by a factor of 1.65 and reduce memory usage. |
Dissecting Transformer Length Extrapolation via the Lens of Receptive Field Analysis (2023.acl-long)
Copied to clipboard
| Challenge: | Length extrapolation allows training a transformer language model on short sequences that preserves perplexities when tested on substantially longer sequences. |
| Approach: | They propose a relative positional embedding design that uses longer than the training sequence to create sandwich. |
| Outcome: | The proposed model can extrapolate to L ex L tr much better than other models. |
ETC: Encoding Long and Structured Inputs in Transformers (2020.emnlp-main)
Copied to clipboard
Joshua Ainslie, Santiago Ontanon, Chris Alberti, Vaclav Cvicek, Zachary Fisher, Philip Pham, Anirudh Ravula, Sumit Sanghai, Qifan Wang, Li Yang
| Challenge: | Existing models for natural language processing (NLP) have been challenging to scale attention to longer inputs. |
| Approach: | They propose an extended Transformer construction architecture that scales attention to longer inputs by combining global-local attention with relative position encodings and a "Contrastive Predictive Coding" objective. |
| Outcome: | The proposed architecture scales attention to longer inputs and encodes structured inputs. |
Attention Alignment and Flexible Positional Embeddings Improve Transformer Length Extrapolation (2024.findings-naacl)
Copied to clipboard
| Challenge: | Existing methods for length extrapolation are tailored for natural language modeling, a task known to have strong recency bias. |
| Approach: | They propose two attention alignment strategies to improve T5's long-context utilization capability without fine-tuning. |
| Outcome: | The proposed methods improve the long-context utilization capability of T5 on language modeling, retrieval, multi-document question answering, and code completion tasks without any fine-tuning. |
Incorporating Noisy Length Constraints into Transformer with Length-aware Positional Encodings (2020.coling-main)
Copied to clipboard
| Challenge: | Neural Machine Translation suffers from an under-translation problem due to limited modeling of output sequence lengths. |
| Approach: | They propose a method to train a Transformer model using length constraints based on positional encoding. |
| Outcome: | The proposed method outperforms a vanilla Transformer in an English-to-Japanese translation by 3.22 points . the noise injection improved robustness for length prediction errors, especially within the window size. |
Context Length Extension via Generalized Extrapolation Scale (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing work on extrapolating positional embedding (RoPE) has limited results in the application of long context language models. |
| Approach: | They propose a set of parameterized extrapolation functions applied to each layer and attention head to adaptively adjust its extrapolations scales. |
| Outcome: | The proposed model achieves stable extrapolation on 64k contexts by training on 16k length text. |