Related papers: TVCACHE: A Stateful Tool-Value Cache for Post-Trai…

Interacting with humans through multi-turn conversations is a fundamental feature of large language models (LLMs). However, existing LLM serving engines executing multi-turn conversations are inefficient due to the need to repeatedly…

Agents utilizing tools powered by large language models (LLMs) or vision-language models (VLMs) have demonstrated remarkable progress in diverse tasks across text and visual modalities. Unlike traditional tools such as calculators, which…

Large Language Models (LLMs) are increasingly used as autonomous agents for multi-step tasks. However, most existing frameworks fail to maintain a structured understanding of the task state, often relying on linear prompt concatenation or…

Vision-Language-Action (VLA) models have demonstrated strong multi-modal reasoning capabilities, enabling direct action generation from visual perception and language instructions in an end-to-end manner. However, their substantial…

A practical large language model (LLM) service may involve a long system prompt, which specifies the instructions, examples, and knowledge documents of the task and is reused across requests. However, the long system prompt causes…

Language models (LMs) are becoming increasingly dependent on external tools. LM-based agentic frameworks frequently interact with their environment via such tools to search files, run code, call APIs, etc. Further, modern reasoning-based…

LLMs are increasingly used world-wide from daily tasks to agentic systems and data analytics, requiring significant GPU resources. LLM inference systems, however, are slow compared to database systems, and inference performance and…

Whether attention key value (KV) states computed for one prompt for a small LLM can be reused to accelerate inference on a new similar prompt, giving an increase to the space to its context memory using an approach called token recycling.…

Skill-distillation pipelines learn reusable rules from LLM agent trajectories, but they lack a key signal: how much each step costs. Without per-step cost, a pipeline cannot distinguish adding a missing step to fix a bug from removing an…

Large-scale deployment of generative AI tools often depends on costly API calls to a Large Language Model (LLM) to fulfil user queries. To curtail the frequency of these calls, one can employ a smaller language model -- a student -- which…

Large language models (LLMs) often incorporate multiple text chunks in their inputs to provide the necessary contexts. To speed up the prefill of the long LLM inputs, one can pre-compute the KV cache of a text and re-use the KV cache when…

Reinforcement Learning (RL) has become the most effective post-training approach for improving the capabilities of Large Language Models (LLMs). In practice, because of the high demands on latency and memory, it is particularly challenging…

Key-Value (KV) cache plays a crucial role in accelerating inference in large language models (LLMs) by storing intermediate attention states and avoiding redundant computation during autoregressive generation. However, its memory footprint…

Caches are used to reduce the speed differential between the CPU and memory to improve the performance of modern processors. However, attackers can use contention-based cache timing attacks to steal sensitive information from victim…

Generating texts with a large language model (LLM) consumes massive amounts of memory. Apart from the already-large model parameters, the key/value (KV) cache that holds information about previous tokens in a sequence can grow to be even…

In Large Language Model (LLM) inference, Key-Value (KV) caches (KV-caches) are essential for reducing time complexity. However, they result in a linear increase in GPU memory as the context length grows. While recent work explores KV-cache…

The key-value (KV) cache is a primary memory bottleneck in Transformers. We propose Low-Rank Key-Value (LRKV) attention, which reduces KV cache memory by exploiting redundancy across attention heads, while being compute efficient. Each…

We consider Time-to-Live (TTL) caches that tag every object in cache with a specific (and possibly renewable) expiration time. State-of-the-art models for TTL caches assume zero object fetch delay, i.e., the time required to fetch a…

Training deep neural networks at the edge on light computational devices, embedded systems and robotic platforms is nowadays very challenging. Continual learning techniques, where complex models are incrementally trained on small batches of…

Video Temporal Grounding (VTG) strives to accurately pinpoint event timestamps in a specific video using linguistic queries, significantly impacting downstream tasks like video browsing and editing. Unlike traditional task-specific models,…