Efficient Length-Generalizable Attention via Causal Retrieval for Long-Context Language Modeling

• URL: http://arxiv.org/abs/2410.01651v2
• Date: Mon, 27 Jan 2025 07:08:49 GMT
• Title: Efficient Length-Generalizable Attention via Causal Retrieval for Long-Context Language Modeling
• Authors: Xiang Hu, Zhihao Teng, Jun Zhao, Wei Wu, Kewei Tu,
• Abstract summary: Grouped Cross Attention can generalize to 1000 times the pre-training context length.<n>Experiments show that GCA-based models achieve near-perfect accuracy in passkey retrieval for 16M context lengths.
• Score: 42.67141329779589
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Despite the success of Transformers, handling long contexts remains challenging due to the limited length generalization and quadratic complexity of self-attention. Thus Transformers often require post-training with a larger attention window, significantly increasing computational and memory costs. In this paper, we propose a novel attention mechanism based on dynamic context, Grouped Cross Attention (GCA), which can generalize to 1000 times the pre-training context length while maintaining the ability to access distant information with a constant attention window size. For a given input sequence, we split it into chunks and use each chunk to retrieve top-k relevant past chunks for subsequent text generation. Specifically, unlike most previous works that use an off-the-shelf retriever, our key innovation allows the retriever to learn how to retrieve past chunks that better minimize the auto-regressive loss of subsequent tokens in an end-to-end manner. Such a mechanism accommodates retrieved chunks with a fixed-size attention window to achieve long-range information access, significantly reducing computational and memory costs during training and inference. Experiments show that GCA-based models achieve near-perfect accuracy in passkey retrieval for 16M context lengths, which is 1000 times the training length.

Related papers

• Squeezed Attention: Accelerating Long Context Length LLM Inference [64.11145320159126]
  We propose Squeezed Attention as a mechanism to accelerate LLM applications where a large portion of the input prompt is fixed. We use K-means clustering offline to group the keys for the fixed context based on semantic similarity and represent each cluster with a single centroid value. We then compute exact attention using only these important keys from the fixed context, thereby reducing bandwidth and computational costs.
  arXiv  Detail & Related papers  (2024-11-14T18:54:19Z)
• Long Sequence Modeling with Attention Tensorization: From Sequence to Tensor Learning [20.51822826798248]
  We propose to scale up the attention field by tensorizing long input sequences into compact tensor representations followed by attention on each transformed dimension. We show that the proposed attention tensorization encodes token dependencies as a multi-hop attention process, and is equivalent to Kronecker decomposition of full attention.
  arXiv  Detail & Related papers  (2024-10-28T11:08:57Z)
• RecurFormer: Not All Transformer Heads Need Self-Attention [14.331807060659902]
  Transformer-based large language models (LLMs) excel in modeling complex language patterns but face significant computational costs during inference. We propose RecurFormer, a novel architecture that replaces certain attention heads with linear recurrent neural networks.
  arXiv  Detail & Related papers  (2024-10-10T15:24:12Z)
• Learning to Retrieve Iteratively for In-Context Learning [56.40100968649039]
  iterative retrieval is a novel framework that empowers retrievers to make iterative decisions through policy optimization. We instantiate an iterative retriever for composing in-context learning exemplars and apply it to various semantic parsing tasks. By adding only 4M additional parameters for state encoding, we convert an off-the-shelf dense retriever into a stateful iterative retriever.
  arXiv  Detail & Related papers  (2024-06-20T21:07:55Z)
• CItruS: Chunked Instruction-aware State Eviction for Long Sequence Modeling [52.404072802235234]
  We introduce Chunked Instruction-aware State Eviction (CItruS), a novel modeling technique that integrates the attention preferences useful for a downstream task into the eviction process of hidden states. Our training-free method exhibits superior performance on long sequence comprehension and retrieval tasks over several strong baselines under the same memory budget.
  arXiv  Detail & Related papers  (2024-06-17T18:34:58Z)
• LongVQ: Long Sequence Modeling with Vector Quantization on Structured Memory [63.41820940103348]
  Self-attention mechanism's computational cost limits its practicality for long sequences. We propose a new method called LongVQ to compress the global abstraction as a length-fixed codebook. LongVQ effectively maintains dynamic global and local patterns, which helps to complement the lack of long-range dependency issues.
  arXiv  Detail & Related papers  (2024-04-17T08:26:34Z)
• Simple and Scalable Strategies to Continually Pre-train Large Language Models [20.643648785602462]
  Large language models (LLMs) are routinely pre-trained on billions of tokens, only to start the process over again once new data becomes available. We show that a simple and scalable combination of learning rate re-warming, LR re-decaying, and replay of previous data is sufficient to match the performance of fully re-training from scratch.
  arXiv  Detail & Related papers  (2024-03-13T17:58:57Z)
• Boot and Switch: Alternating Distillation for Zero-Shot Dense Retrieval [50.47192086219752]
  $texttABEL$ is a simple but effective unsupervised method to enhance passage retrieval in zero-shot settings. By either fine-tuning $texttABEL$ on labelled data or integrating it with existing supervised dense retrievers, we achieve state-of-the-art results.
  arXiv  Detail & Related papers  (2023-11-27T06:22:57Z)
• MiLe Loss: a New Loss for Mitigating the Bias of Learning Difficulties in Generative Language Models [40.992566245706996]
  We propose a MiLe Loss function for mitigating the bias of learning difficulties with tokens. We train generative language models at different scales of 468M, 1.2B, and 6.7B parameters. Experiments reveal that models incorporating the proposed MiLe Loss can gain consistent performance improvement on downstream benchmarks.
  arXiv  Detail & Related papers  (2023-10-30T13:33:21Z)
• Ring Attention with Blockwise Transformers for Near-Infinite Context [88.61687950039662]
  We present a novel approach, Ring Attention with Blockwise Transformers (Ring Attention), which leverages blockwise computation of self-attention and feedforward to distribute long sequences across multiple devices. Our approach enables training and inference of sequences that are up to device count times longer than those achievable by prior memory-efficient Transformers.
  arXiv  Detail & Related papers  (2023-10-03T08:44:50Z)
• Focused Transformer: Contrastive Training for Context Scaling [31.44508996359732]
  We introduce the Focused Transformer (FoT), a technique that employs a training process inspired by contrastive learning. FoT enhances the structure of the (key, value) space, enabling an extension of the context length. Our method allows for fine-tuning pre-existing, large-scale models to lengthen their effective context.
  arXiv  Detail & Related papers  (2023-07-06T17:52:10Z)
• LongNet: Scaling Transformers to 1,000,000,000 Tokens [146.4077038371075]
  LongNet is a Transformer variant that can scale sequence length to more than 1 billion tokens. Our work opens up new possibilities for modeling very long sequences, e.g., treating a whole corpus or even the entire Internet as a sequence.
  arXiv  Detail & Related papers  (2023-07-05T17:59:38Z)
• Retrieval-Pretrained Transformer: Long-range Language Modeling with Self-retrieval [51.437420003471615]
  We propose the Retrieval-Pretrained Transformer (RPT), an architecture and training procedure for jointly training a retrieval-augmented LM from scratch. RPT improves retrieval quality and subsequently perplexity across the board compared to strong baselines.
  arXiv  Detail & Related papers  (2023-06-23T10:18:02Z)
• Generate to Understand for Representation [3.5325087487696463]
  GUR is a pretraining framework that combines language modeling and contrastive learning objectives in a single training step. GUR achieves impressive results without any labeled training data, outperforming all other pretrained baselines as a retriever at the recall benchmark in a zero-shot setting.
  arXiv  Detail & Related papers  (2023-06-14T06:00:18Z)
• Landmark Attention: Random-Access Infinite Context Length for Transformers [45.69864961773124]
  We present a novel approach that allows access to the complete context while retaining random-access flexibility. Our method uses a landmark token to represent each block of the input and trains the attention to use it for selecting relevant blocks. We demonstrate that our method can obtain comparable performance with Transformer-XL while significantly reducing the number of retrieved tokens in each step.
  arXiv  Detail & Related papers  (2023-05-25T17:53:42Z)
• Scaling Transformer to 1M tokens and beyond with RMT [5.60052250541419]
  A major limitation for the broader scope of problems solvable by transformers is the quadratic scaling of computational complexity with input size. In this study, we investigate the recurrent memory augmentation of pre-trained transformer models to extend input context length while linearly scaling compute. Our approach demonstrates the capability to store information in memory for sequences of up to an unprecedented two million tokens while maintaining high retrieval accuracy.
  arXiv  Detail & Related papers  (2023-04-19T16:18:54Z)
• Characterizing Attribution and Fluency Tradeoffs for Retrieval-Augmented Large Language Models [6.425088990363101]
  We examine the relationship between fluency and attribution in Large Language Models prompted with retrieved evidence. We show that larger models tend to do much better in both fluency and attribution. We propose a recipe that could allow smaller models to both close the gap with larger models and preserve the benefits of top-k retrieval.
  arXiv  Detail & Related papers  (2023-02-11T02:43:34Z)
• Language Model Pre-Training with Sparse Latent Typing [66.75786739499604]
  We propose a new pre-training objective, Sparse Latent Typing, which enables the model to sparsely extract sentence-level keywords with diverse latent types. Experimental results show that our model is able to learn interpretable latent type categories in a self-supervised manner without using any external knowledge.
  arXiv  Detail & Related papers  (2022-10-23T00:37:08Z)
• Unsupervised Paraphrasing with Pretrained Language Models [85.03373221588707]
  We propose a training pipeline that enables pre-trained language models to generate high-quality paraphrases in an unsupervised setting. Our recipe consists of task-adaptation, self-supervision, and a novel decoding algorithm named Dynamic Blocking. We show with automatic and human evaluations that our approach achieves state-of-the-art performance on both the Quora Question Pair and the ParaNMT datasets.
  arXiv  Detail & Related papers  (2020-10-24T11:55:28Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.