Gated Differentiable Working Memory for Long-Context Language Modeling

• URL: http://arxiv.org/abs/2601.12906v1
• Date: Mon, 19 Jan 2026 10:00:33 GMT
• Title: Gated Differentiable Working Memory for Long-Context Language Modeling
• Authors: Lingrui Mei, Shenghua Liu, Yiwei Wang, Yuyao Ge, Baolong Bi, Jiayu Yao, Jun Wan, Ziling Yin, Jiafeng Guo, Xueqi Cheng,
• Abstract summary: We propose Gdwm (Gated Differentiable Working Memory), a framework that introduces a write controller to gate the consolidation process.<n>Experiments on ZeroSCROLLS and LongBench v2 demonstrate that Gdwm achieves comparable or superior performance with 4$times$ fewer gradient steps than uniform baselines.
• Score: 80.27483324685434
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Long contexts challenge transformers: attention scores dilute across thousands of tokens, critical information is often lost in the middle, and models struggle to adapt to novel patterns at inference time. Recent work on test-time adaptation addresses this by maintaining a form of working memory -- transient parameters updated on the current context -- but existing approaches rely on uniform write policies that waste computation on low-utility regions and suffer from high gradient variance across semantically heterogeneous contexts. In this work, we reframe test-time adaptation as a budget-constrained memory consolidation problem, focusing on which parts of the context should be consolidated into working memory under limited computation. We propose Gdwm (Gated Differentiable Working Memory), a framework that introduces a write controller to gate the consolidation process. The controller estimates Contextual Utility, an information-theoretic measure of long-range contextual dependence, and allocates gradient steps accordingly while maintaining global coverage. Experiments on ZeroSCROLLS and LongBench v2 demonstrate that Gdwm achieves comparable or superior performance with 4$\times$ fewer gradient steps than uniform baselines, establishing a new efficiency-performance Pareto frontier for test-time adaptation.

Related papers

• TempoNet: Slack-Quantized Transformer-Guided Reinforcement Scheduler for Adaptive Deadline-Centric Real-Time Dispatchs [8.818252253980985]
  TempoNet is a reinforcement learning scheduler that pairs a permutation-invariant Transformer with a deep Q-approximation.<n>A latency-aware sparse attention stack with blockwise top-k selection and locality-sensitive chunking enables global reasoning over unordered task sets.
  arXiv  Detail & Related papers  (2026-02-20T09:56:23Z)
• Rethinking Multi-Condition DiTs: Eliminating Redundant Attention via Position-Alignment and Keyword-Scoping [61.459927600301654]
  Multi-condition control is bottlenecked by the conventional concatenate-and-attend'' strategy.<n>Our analysis reveals that much of this cross-modal interaction is spatially or semantically redundant.<n>We propose Position-aligned and Keyword-scoped Attention (PKA), a highly efficient framework designed to eliminate these redundancies.
  arXiv  Detail & Related papers  (2026-02-06T16:39:10Z)
• Training-Free Test-Time Adaptation with Brownian Distance Covariance in Vision-Language Models [16.03043781097689]
  Training-free Test-Time Adaptation with Brownian Distance Covariance (TaTa)<n>TaTa leverages Brownian Distance Covariance to dynamically adapt vision-language models to new domains without training or back-propagation.<n>Experiments across diverse datasets show that TaTa significantly reduces computational cost while achieving state-of-the-art performance in domain and cross-dataset generalization.
  arXiv  Detail & Related papers  (2026-01-30T18:21:45Z)
• AMA: Adaptive Memory via Multi-Agent Collaboration [54.490349689939166]
  We propose Adaptive Memory via Multi-Agent Collaboration (AMA), a novel framework that leverages coordinated agents to manage memory across multiple granularities.<n>AMA significantly outperforms state-of-the-art baselines while reducing token consumption by approximately 80% compared to full-context methods.
  arXiv  Detail & Related papers  (2026-01-28T08:09:49Z)
• Gated KalmaNet: A Fading Memory Layer Through Test-Time Ridge Regression [53.48692193399171]
  Gated KalmaNet (GKA) is a layer that reduces the gap by accounting for the full past when predicting the next token.<n>We solve an online ridge regression problem at test time, with constant memory and linear compute cost in the sequence length.<n>On long-context, GKA excels at real-world RAG and LongQA tasks up to 128k tokens, achieving more than $10$% relative improvement over other fading memory baselines.
  arXiv  Detail & Related papers  (2025-11-26T03:26:37Z)
• PPMStereo: Pick-and-Play Memory Construction for Consistent Dynamic Stereo Matching [51.98089287914147]
  textbfPick-and-textbflay textbfMemory (PM) construction module for dynamic bfStereo matching, dubbed as bftextPPMStereo.<n>Inspired by the two-stage decision-making process in humans, we propose a textbfPick-and-textbflay textbfMemory (PM) construction module for dynamic bfStereo matching, dubbed as bftextPPMStereo.
  arXiv  Detail & Related papers  (2025-10-23T03:52:39Z)
• ResFormer: All-Time Reservoir Memory for Long Sequence Classification [4.298381633106637]
  Sequence classification is essential in NLP for understanding and categorizing language patterns in tasks like sentiment analysis, intent detection, and topic classification.<n> Transformer-based models, despite achieving state-of-the-art performance, have inherent limitations due to quadratic time and memory complexity.<n>We propose ResFormer, a novel neural network architecture designed to model varying context lengths efficiently through a cascaded methodology.
  arXiv  Detail & Related papers  (2025-09-28T21:20:49Z)
• On-the-Fly Adaptive Distillation of Transformer to Dual-State Linear Attention [53.22963042513293]
  Large language models (LLMs) excel at capturing global token dependencies via self-attention but face prohibitive compute and memory costs on lengthy inputs.<n>We first propose dual-state linear attention (A), a novel design that maintains two hidden states-one for preserving historical context and one for tracking recencythereby mitigating the short-range bias typical of linear-attention architectures.<n>We introduce DSLA-Serve, an online adaptive distillation framework that progressively replaces Transformer layers DSLA layers at inference time, guided by a sensitivity-based layer ordering.
  arXiv  Detail & Related papers  (2025-06-11T01:25:06Z)
• Compress, Gather, and Recompute: REFORMing Long-Context Processing in Transformers [58.98923344096319]
  REFORM is a novel inference framework that efficiently handles long contexts through a two-phase approach.<n>It achieves over 50% and 27% performance gains on RULER and BABILong respectively at 1M context length.<n>It also outperforms baselines on Infinite-Bench and MM-NIAH, demonstrating flexibility across diverse tasks and domains.
  arXiv  Detail & Related papers  (2025-06-01T23:49:14Z)
• ReservoirTTA: Prolonged Test-time Adaptation for Evolving and Recurring Domains [25.075869018443925]
  ReservoirTTA is a novel plug-in framework designed for prolonged test-time adaptation.<n>At its core, ReservoirTTA maintains a reservoir of domain-specialized models.<n>Our theoretical analysis reveals key components that bound parameter variance and prevent model collapse.
  arXiv  Detail & Related papers  (2025-05-20T15:39:20Z)
• Contextual Memory Reweaving in Large Language Models Using Layered Latent State Reconstruction [0.0]
  Token dependencies degrade as sequence length increases, leading to a decline in coherence and factual consistency.<n>A structured approach is introduced to mitigate this issue through the reweaving of latent states captured at different processing layers.<n>The proposed Contextual Memory Reweaving framework incorporates a Layered Latent State Reconstruction mechanism.
  arXiv  Detail & Related papers  (2025-02-04T06:25:20Z)
• On the Convergence of Continual Federated Learning Using Incrementally Aggregated Gradients [7.226144684379189]
  The holy grail of machine learning is to enable Continual Federated Learning (CFL) to enhance the efficiency, privacy, and scalability of AI systems while learning from streaming data.<n>We propose a novel replay-memory based federated strategy consisting of edge-based gradient updates on memory and aggregated gradients on the current data.<n>We empirically show that C-FLAG outperforms several state-of-the-art baselines on both task and class-incremental settings with respect to metrics such as accuracy and forgetting.
  arXiv  Detail & Related papers  (2024-11-12T17:36:20Z)
• SHERL: Synthesizing High Accuracy and Efficient Memory for Resource-Limited Transfer Learning [63.93193829913252]
  We propose an innovative METL strategy called SHERL for resource-limited scenarios. In the early route, intermediate outputs are consolidated via an anti-redundancy operation. In the late route, utilizing minimal late pre-trained layers could alleviate the peak demand on memory overhead.
  arXiv  Detail & Related papers  (2024-07-10T10:22:35Z)

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.