This work employs encoder-decoder networks built around long-short term memory (LSTM) cells in a distributed optimization algorithm that allows unprecedented reconstruction accuracy or the usage of different prediction-horizon lengths without the necessity to retrain.
Abstract
The communication demands of distributed model prediction control (DMPC) can overwhelm even advanced wireless communication technologies as agents must exchange a significant amount of information at least once per time step. To semantically reduce communication demands, this work employs encoder-decoder networks built around long-short term memory (LSTM) cells in a distributed optimization algorithm. Agents publish a reduced representation of a message and receivers reconstruct the original message upon reception. In tests with reduced communication using formations of mobile robots, trained networks retain satisfactory performance and work reliably under conditions overwhelming full communication. As the results show, the usage of LSTMs either allows unprecedented reconstruction accuracy or the usage of different prediction-horizon lengths without the necessity to retrain.
Retrieval-augmented experience replay is an emerging approach for the parameter-free adaptation of large language model (LLM) agents at inference time. However, injecting raw multi-step trajectories into the agent's context is costly and risks flooding the prompt with irrelevant details as the experience pool grows. We propose Structured Experience Compression (SEC), a lightweight method that distills each trajectory into a four-field unit encoding the task type $(T)$, key decisions $(D)$, failure reason $(F)$, and repair action $(R)$. At inference time the top- $k$ units closest to the current task are retrieved and appended to the prompt. Experiments on ALFWorld and WebShop show that SEC outperforms rawtrajectory retrieval by 7.7 percentage points on ALFWorld and by an absolute reward gain of 0.056 on WebShop, while consuming only 37% of the context tokens. SEC also degrades more gracefully when the pool contains noisy or mismatched records, suggesting potential suitability for long-lived deployment settings.
Zhizhen Zhu, Ruochen Fang, Rong Hu· 2026 7th International Confe...· 1 citation
Deploying large language models in resource-constrained environments such as mobile devices requires effective compression. Pruning and quantization are largely used, yet applying pruning to a new model typically demands manual inspection of its architecture to identify layer names, MLP structure, and projection patterns, a process that is tedious, error-prone, and must be repeated for every new model design. We present ShrinkLLM, an end-to-end compression system that eliminates the manual inspection of model structure during pruning by delegating architectural analysis to an AI agent. Given a causal LLM, the agent automatically identifies MLP layer organization, expansion layers, gate projections for Gated Linear Unit architectures, and down-projection layers, structural information that is then passed directly to a compression engine to perform neuron pruning, quantization, or a combination of both. We demonstrate the system on llama-3.1-8b-instruct, Qwen2.5-3b-instruct and Gemma-2-2b-it showcasing how the agent correctly resolves architecturespecific structural details and drives compression across all MLP layers automatically.
Gyanko Issah Yussif, Tasnia Sultana, Mohamed Ali et al.· International Conference on...· 0 citations
Large Language Models (LLMs) trained using Chain-of-Thought (CoT) supervision have achieved state-of-the-art performance on complex reasoning tasks. However, the generation of long reasoning chains introduces substantial computational overhead during inference, limiting their deployment in low-latency and resource-constrained environments. This paper proposes AdaptiReason, a novel framework that dynamically compresses intermediate reasoning steps based on task difficulty and model confidence without requiring retraining of the underlying base model. AdaptiReason employs a lightweight difficulty estimator to determine the appropriate reasoning depth for each input, followed by a learned token-pruning policy that eliminates redundant or low-information reasoning steps. Experimental evaluation on the MATH, GSM8K, and ARC-Challenge benchmarks demonstrates that AdaptiReason reduces the average number of generated tokens by 3.7× while preserving 98.2% of the baseline reasoning accuracy. Furthermore, the proposed framework is model-agnostic and can be seamlessly integrated with instruction-tuned LLMs without requiring access to model parameters, relying solely on output logits for adaptive reasoning compression. The results demonstrate that AdaptiReason significantly improves inference efficiency while maintaining high reasoning performance, making it suitable for real-time and resource-constrained LLM applications.
V. A, Mithaguru, Amrita Kundu et al.· 2026 4th International Confe...· 0 citations
Modern LLM coding agents such as Claude Code and OpenHands share a common inefficiency: they spend much of their token budget finding the file to patch, rather than patching it. On SWE-Bench Verified, a 30B OpenHands agent averages 23 rounds and 631K tokens per resolved issue, with many calls spent on grep, glob, and view_file during repository exploration. We introduce CodeGrep, a 14B retrieval agent trained end-to-end with GRPO to issue multi-turn parallel grep, glob, and read tool calls and return candidate files to a frozen downstream coding agent. On all 500 SWE-Bench Verified instances, CodeGrep preserves resolve rate while substantially improving efficiency: 27.0% versus 25.8% for the no-retrieval baseline, with 15% fewer rounds and 19% fewer tokens on resolved instances. Across retrievers, downstream utility follows a precision threshold: BM25 with precision 0.375 degrades the agent, Jina with precision 0.445 is neutral, and CodeGrep with precision 0.677 crosses the threshold at which retrieval begins to reduce rollout cost. To enable this study, we mine supervision from 67K open-source agent trajectories using CATM and build a Git-worktree environment for multi-turn agent RL. In our setting, applying the efficiency signal at the advantage layer rather than the reward layer reduces KL drift and translates cleanly into downstream efficiency. We will release the model, training pipeline, RL environment, and evaluation harnesses.
Wuya Chen, Yihao Yang, Yang Cao et al.· 0 citations
Steerable pluralistic alignment aims to enable large language models (LLMs) to reliably adhere to diverse and potentially conflicting human values, particularly when target objectives involve multi-dimensional, compositional values. Current methods largely rely on prompt engineering or reasoning-time guidance, which often results in fragile and non-persistent con-trol once prompts are perturbed or omitted. In this work, we study value-controllable alignment through discrete condition vectors and propose Verifiable-reward-Routed LoRA—a parameter-efficient mixture-of-experts LoRA framework enhanced with conditioned gating. This gating mechanism dynamically directs the flow among multiple LoRA experts based on an input value or moral vector. To ensure that such routing leads to semantically compliant outputs, we formulate post-training as a reinforcement learning problem with verifiable rewards. We further introduce a conditional consistency reward, computed by an external model-based verifier implemented as a lightweight discriminator, and optimize the adapter parameters using GRPO. Experiments on the Touché23-valueEval (value alignment) and MIC (moral alignment) benchmarks, using two 8-billion-parameter back-bones, show that our method consistently out-performs prompt-based steering and multi-task PEFT baselines. It attains the highest over-all controllability across micro-F1, macro-F1, and Jaccard metrics—a conclusion further reinforced by human pairwise evaluations
Jing Wang, Yaomin Wu, Yinglin Wang et al.· Annual Meeting of the Associ...· 0 citations
LLM-based coding agents solve software-engineering tasks through iterative interactions with development environments, where returned observations accumulate in the context and become a major source of inference cost. Observation compression reduces this cost by shortening observations before they are appended to the context. However, existing methods still exhibit an unsatisfactory efficiency-effectiveness trade-off, as they do not explicitly model how compression affects the agent's subsequent behavior. This paper proposes CoACT, an action-preserving observation compression method for coding agents. CoACT is built on next-action preservation (NAP), which requires a compressed observation to induce the same next action as the raw observation. By checking the agent's immediate next action, NAP provides a practical signal for whether a compression preserves the information needed for continued task solving. During training, a teacher model first generates multiple compressed candidates of each observation. CoACT then uses an action-preservation reward based on NAP to filter out candidates that would change the agent's next action, and uses a length-reduction reward to choose compact candidates as supervision for a lightweight compressor. Experiments on SWE-bench Verified with three agentic models show that CoACT reduces average total token consumption by 33.0% while maintaining task-solving effectiveness close to the uncompressed agent.
Hao Chen, Yuancheng Zhu, Yitong Zhang et al.· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduAug 18, 2026
A new method for surgically removing training examples from a model reveals that as datasets grow, the link between what a model learns and what it produces dissolves.
