PageSourceSearch

https://vllm-sr.ai/publications/

html vllm-sr.ai collected 2026-10-04 01:30:31 UTC 30,973 bytes, 31 lines download raw bytes

1<!doctype html>
2<html lang="en-US" dir="ltr" class="plugin-pages plugin-id-default" data-has-hydrated="false">
3<head>
4<meta charset="UTF-8">
5<meta name="generator" content="Docusaurus v3.10.2">
6<title data-rh="true">Papers &amp; Talks | vLLM Semantic Router</title><meta data-rh="true" name="twitter:url" content="https://vllm-sr.ai/publications"><meta data-rh="true" name="viewport" content="width=device-width,initial-scale=1"><meta data-rh="true" property="og:url" content="https://vllm-sr.ai/publications"><meta data-rh="true" property="og:locale" content="en_US"><meta data-rh="true" property="og:locale:alternate" content="zh_Hans"><meta data-rh="true" name="docusaurus_locale" content="en"><meta data-rh="true" name="docusaurus_tag" content="default"><meta data-rh="true" name="docsearch:language" content="en"><meta data-rh="true" name="docsearch:docusaurus_tag" content="default"><meta data-rh="true" name="keywords" content="Mixture-of-Models, heterogeneous LLM inference, preference-driven AI, open-source LLM router, multi-model routing, model orchestration, model selection, model cascade, Fusion API, semantic router, vLLM"><meta data-rh="true" name="author" content="vLLM Semantic Router Team"><meta data-rh="true" name="application-name" content="vLLM Semantic Router"><meta data-rh="true" property="og:type" content="website"><meta data-rh="true" property="og:site_name" content="vLLM Semantic Router"><meta data-rh="true" property="og:image" content="https://vllm-sr.ai/img/vllm-sr-logo.social.png"><meta data-rh="true" property="og:image:alt" content="vLLM Semantic Router social preview"><meta data-rh="true" name="twitter:card" content="summary_large_image"><meta data-rh="true" name="twitter:title" content="Mixture-of-Models for Heterogeneous LLM Inference | vLLM Semantic Router"><meta data-rh="true" name="twitter:description" content="Mixture-of-Models is a serving-system architecture for heterogeneous LLM inference. vLLM Semantic Router makes it executable."><meta data-rh="true" name="twitter:image" content="https://vllm-sr.ai/img/vllm-sr-logo.social.png"><meta data-rh="true" name="twitter:image:alt" content="vLLM Semantic Router social preview"><meta data-rh="true" name="geo.region" content="US-CA"><meta data-rh="true" name="geo.placename" content="San Francisco"><meta data-rh="true" name="geo.position" content="37.7749;-122.4194"><meta data-rh="true" name="ICBM" content="37.7749, -122.4194"><meta data-rh="true" property="og:title" content="Papers &amp; Talks | vLLM Semantic Router"><meta data-rh="true" name="description" content="Latest research publications, talks, and scientific contributions from the vLLM Semantic Router project."><meta data-rh="true" property="og:description" content="Latest research publications, talks, and scientific contributions from the vLLM Semantic Router project."><link data-rh="true" rel="icon" href="/img/vllm.png"><link data-rh="true" rel="canonical" href="https://vllm-sr.ai/publications"><link data-rh="true" rel="alternate" href="https://vllm-sr.ai/publications" hreflang="en-US"><link data-rh="true" rel="alternate" href="https://vllm-sr.ai/zh-Hans/publications" hreflang="zh-Hans"><link data-rh="true" rel="alternate" href="https://vllm-sr.ai/publications" hreflang="x-default"><link rel="alternate" type="application/rss+xml" href="/blog/rss.xml" title="vLLM Semantic Router RSS Feed">
7<link rel="alternate" type="application/atom+xml" href="/blog/atom.xml" title="vLLM Semantic Router Atom Feed">
8
9
10
11
12
13
14
15
16<link rel="preconnect" href="https://fonts.googleapis.com">
17<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin="anonymous">
18<link rel="stylesheet" href="https://fonts.googleapis.com/css2?family=Inter:opsz,[email protected],400..700&amp;
18family=JetBrains+Mono:wght@400;500;600&amp;display=swap">
19<script type="application/ld+json">{"@context":"https://schema.org","@type":"SoftwareApplication","name":"vLLM Semantic Router","applicationCategory":"AIInfrastructure","operatingSystem":"Cross-platform","description":"Mixture-of-Models is a serving-system architecture for heterogeneous LLM inference. vLLM Semantic Router makes it executable.","url":"https://vllm-sr.ai","publisher":{"@type":"Organization","name":"vLLM Semantic Router Team","url":"https://github.com/vllm-project/semantic-router"}}</script>
19<link rel="stylesheet" href="/assets/css/styles.d108128e.css">
20<script src="/assets/js/runtime~main.0f68cf72.js" defer="defer"></script>
vendor: 1 bytes, line 20
20
21<script src="/assets/js/main.c286df1d.js" defer="defer"></script>
21
22</head>
23<body>
24<svg style="display: none;"><defs>
25<symbol id="theme-svg-external-link" viewBox="0 0 24 24"><path fill="currentColor" d="M21 13v10h-21v-19h12v2h-10v15h17v-8h2zm3-12h-10.988l4.035 4-6.977 7.07 2.828 2.828 6.977-7.07 4.125 4.172v-11z"/></symbol>
26</defs></svg>
27<script>!function(){var t=function(){try{return new URLSearchParams(window.location.search).get("docusaurus-theme")}catch(t){}}()||function(){try{return window.localStorage.getItem("theme")}catch(t){}}();document.documentElement.setAttribute("data-theme",t||(window.matchMedia("(prefers-color-scheme: dark)").matches?"dark":"light")),document.documentElement.setAttribute("data-theme-choice",t||"system")}(),function(){try{const c=new URLSearchParams(window.location.search).entries();for(var[t,e]of c)if(t.startsWith("docusaurus-data-")){var a=t.replace("docusaurus-data-","data-");document.documentElement.setAttribute(a,e)}}catch(t){}}()</script>
27<div id="__docusaurus"><div class="site-root site-root--page site-page--publications"><div role="region" aria-label="Skip to main content"><a class="skipToContent_fXgn" href="#__docusaurus_skipToContent_fallback">Skip to main content</a></div><nav aria-label="Main" class="theme-layout-navbar navbar navbar--fixed-top"><div class="navbar__inner"><div class="theme-layout-navbar-left navbar__items"><button aria-label="Toggle navigation bar" aria-expanded="false" class="navbar__toggle clean-btn" type="button"><svg width="30" height="30" viewBox="0 0 30 30" aria-hidden="true"><path stroke="currentColor" stroke-linecap="round" stroke-miterlimit="10" stroke-width="2" d="M4 7h22M4 15h22M4 23h22"></path></svg></button><a class="navbar__brand" href="/"><div class="navbar__logo"><img src="/img/vllm-sr-logo.light.png" alt="vLLM Semantic Router Logo" class="themedComponent_mlkZ themedComponent--light_NVdE"><img src="/img/vllm-sr-logo.white.png" alt="vLLM Semantic Router Logo" class="themedComponent_mlkZ themedComponent--dark_xIcU"></div></a><nav class="root__O1p" aria-label="Primary navigation"><a class="link_lE5P" href="/docs/intro">Docs</a><a class="link_lE5P linkActive_sV4t" href="/publications">Research</a><a class="link_lE5P" href="/blog">Blog</a><a class="link_lE5P" href="/community/team">Community</a></nav></div><div class="theme-layout-navbar-right navbar__items navbar__items--right"><div class="navbarSearchContainer_Bca1"><button aria-haspopup="dialog" aria-label="Search the docs and blog…" class="site-search-trigger" type="button"><svg aria-hidden="true" fill="none" height="15" stroke="currentColor" stroke-linecap="round" stroke-width="2" viewBox="0 0 24 24" width="15"><circle cx="11" cy="11" r="7"></circle><path d="M20 20l-3.5-3.5"></path></svg><span class="site-search-trigger__label">Search</span><kbd class="site-search-trigger__key">Ctrl K</kbd></button></div><div class="navbar__item dropdown dropdown--hoverable dropdown--right"><a href="#" aria-haspopup="true" aria-expanded="false" role="button" class="navbar__link nav-locale"><svg viewBox="0 0 24 24" width="20" height="20" aria-hidden="true" class="iconLanguage_DSK9"><path fill="currentColor" d="M12.87 15.07l-2.54-2.51.03-.03c1.74-1.94 2.98-4.17 3.71-6.53H17V4h-7V2H8v2H1v1.99h11.17C11.5 7.92 10.44 9.75 9 11.35 8.07 10.32 7.3 9.19 6.69 8h-2c.73 1.63 1.73 3.17 2.98 4.56l-5.09 5.02L4 19l5-5 3.11 3.11.76-2.04zM18.5 10h-2L12 22h2l1.12-3h4.75L21 22h2l-4.5-12zm-2.62 7l1.62-4.33L19.12 17h-3.24z"></path></svg>English</a><ul class="dropdown__menu"><li><a href="/publications" target="_self" rel="noopener noreferrer" class="dropdown__link dropdown__link--active" lang="en-US">English</a></li><li><a href="/zh-Hans/publications" target="_self" rel="noopener noreferrer" class="dropdown__link" lang="zh-Hans">简体中文</a></li></ul></div><div class="navbar__item dropdown dropdown--hoverable dropdown--right"><a class="navbar__link nav-docs-only" aria-haspopup="true" aria-expanded="false" role="button" href="/docs/intro">Latest</a><ul class="dropdown__menu"><li><a class="dropdown__link" href="/docs/intro">Latest</a></li><li><a class="dropdown__link" href="/docs/v0.3/intro">v0.3</a></li><li><a class="dropdown__link" href="/docs/v0.2/intro">v0.2</a></li><li><a class="dropdown__link" href="/docs/v0.1/intro">v0.1</a></li></ul></div><div class="navbar__item dropdown dropdown--hoverable dropdown--right"><a href="#" aria-haspopup="true" aria-expanded="false" role="button" class="navbar__link nav-utility">More</a><ul class="dropdown__menu"><li><a class="dropdown__link" href="/models">Model Hub</a></li><li><a class="dropdown__link" href="/community/contributors">Leaderboard</a></li><li><a href="https://github.com/vllm-project/semantic-router" target="_blank" rel="noopener noreferrer" class="dropdown__link">GitHub<svg width="12" height="12" aria-label="(opens in new tab)" class="iconExternalLink_nPIU"><use href="#theme-svg-external-link"></use></svg></a></li><li><a href="https://huggingface.co/vllm-sr" target="_blank" rel="noopener noreferrer" class="dropdown__link">Hugging Face<svg width="12" height="12" aria-label="(opens in new tab)" class="iconExternalLink_nPIU"><use href="#theme-svg-external-link"></use></svg></a></li><li><a href="https://github.com/vllm-project/semantic-router/discussions" target="_blank" rel="noopener noreferrer" class="dropdown__link">Discussions<svg width="12" height="12" aria-label="(opens in new tab)" class="iconExternalLink_nPIU"><use href="#theme-svg-external-link"></use></svg></a></li></ul></div><div class="toggle_vylO colorModeToggle_x44X navbar-color-mode-toggle"><button class="clean-btn toggleButton_gllP toggleButtonDisabled_aARS" type="button" disabled="" title="light mode" aria-label="Switch between dark and light mode (currently light mode)"><svg viewBox="0 0 24 24" width="24" height="24" aria-hidden="true" class="toggleIcon_g3eP lightToggleIcon_pyhR">
27<path fill="currentColor" d="M12,9c1.65,0,3,1.35,3,3s-1.35,3-3,3s-3-1.35-3-3S10.35,9,12,9 M12,7c-2.76,0-5,2.24-5,5s2.24,5,5,5s5-2.24,5-5 S14.76,7,12,7L12,7z M2,13l2,0c0.55,0,1-0.45,1-1s-0.45-1-1-1l-2,0c-0.55,0-1,0.45-1,1S1.45,13,2,13z M20,13l2,0c0.55,0,1-0.45,1-1 s-0.45-1-1-1l-2,0c-0.55,0-1,0.45-1,1S19.45,13,20,13z M11,2v2c0,0.55,0.45,1,1,1s1-0.45,1-1V2c0-0.55-0.45-1-1-1S11,1.45,11,2z M11,20v2c0,0.55,0.45,1,1,1s1-0.45,1-1v-2c0-0.55-0.45-1-1-1C11.45,19,11,19.45,11,20z M5.99,4.58c-0.39-0.39-1.03-0.39-1.41,0 c-0.39,0.39-0.39,1.03,0,1.41l1.06,1.06c0.39,0.39,1.03,0.39,1.41,0s0.39-1.03,0-1.41L5.99,4.58z M18.36,16.95 c-0.39-0.39-1.03-0.39-1.41,0c-0.39,0.39-0.39,1.03,0,1.41l1.06,1.06c0.39,0.39,1.03,0.39,1.41,0c0.39-0.39,0.39-1.03,0-1.41 L18.36,16.95z M19.42,5.99c0.39-0.39,0.39-1.03,0-1.41c-0.39-0.39-1.03-0.39-1.41,0l-1.06,1.06c-0.39,0.39-0.39,1.03,0,1.41 s1.03,0.39,1.41,0L19.42,5.99z M7.05,18.36c0.39-0.39,0.39-1.03,0-1.41c-0.39-0.39-1.03-0.39-1.41,0l-1.06,1.06 c-0.39,0.39-0.39,1.03,0,1.41s1.03,0.39,1.41,0L7.05,18.36z"></path></svg><svg viewBox="0 0 24 24" width="24" height="24" aria-hidden="true" class="toggleIcon_g3eP darkToggleIcon_wfgR"><path fill="currentColor" d="M9.37,5.51C9.19,6.15,9.1,6.82,9.1,7.5c0,4.08,3.32,7.4,7.4,7.4c0.68,0,1.35-0.09,1.99-0.27C17.45,17.19,14.93,19,12,19 c-3.86,0-7-3.14-7-7C5,9.07,6.81,6.55,9.37,5.51z M12,3c-4.97,0-9,4.03-9,9s4.03,9,9,9s9-4.03,9-9c0-0.46-0.04-0.92-0.1-1.36 c-0.98,1.37-2.58,2.26-4.4,2.26c-2.98,0-5.4-2.42-5.4-5.4c0-1.81,0.89-3.42,2.26-4.4C12.92,3.04,12.46,3,12,3L12,3z"></path></svg><svg viewBox="0 0 24 24" width="24" height="24" aria-hidden="true" class="toggleIcon_g3eP systemToggleIcon_QzmC"><path fill="currentColor" d="m12 21c4.971 0 9-4.029 9-9s-4.029-9-9-9-9 4.029-9 9 4.029 9 9 9zm4.95-13.95c1.313 1.313 2.05 3.093 2.05 4.95s-0.738 3.637-2.05 4.95c-1.313 1.313-3.093 2.05-4.95 2.05v-14c1.857 0 3.637 0.737 4.95 2.05z"></path></svg></button></div><a href="https://app.vllm-sr.ai" target="_blank" rel="noopener noreferrer" class="navbar__item navbar__link nav-dashboard-cta">Dashboard<svg width="13.5" height="13.5" aria-label="(opens in new tab)" class="iconExternalLink_nPIU"><use href="#theme-svg-external-link"></use></svg></a></div></div><div role="presentation" class="navbar-sidebar__backdrop"></div></nav><div id="__docusaurus_skipToContent_fallback" class="theme-layout-main main-wrapper mainWrapper_z2l0"><div class="page_cI6e"><main class="container_fwUJ"><header class="site-page-header header_HjTW"><div class="text_wspW"><span class="eyebrow_TNoS">Research</span><h1 class="title_svgD">Papers &amp; Talks</h1><p class="description_pnrv">Research, talks, and position papers from the vLLM Semantic Router project.</p></div></header><div class="body_IQfd"><nav class="sidebar_ekPy" aria-label="Research sections"><div class="navGroup_cyWD"><span class="navGroupLabel_VbBW">Publications</span><a class="navLink_REva navLinkActive_aDvp" aria-current="page" href="/publications">Papers &amp; Talks</a><a class="navLink_REva" href="/white-paper">White Paper</a><a class="navLink_REva" href="/vision-paper">Vision Paper</a><a class="navLink_REva" href="/decision-paper">Decision Paper</a></div></nav><article class="article_DsnW"><div class="filterSection_gCNC"><div class="filterButtons_aJt8"><button class="filterButton_O8L5 active_I77t">All<!-- --> <!-- -->(<!-- -->22<!-- -->)</button><button class="filterButton_O8L5">Papers<!-- --> <!-- -->(<!-- -->19<!-- -->)</button><button class="filterButton_O8L5">Talks<!-- --> <!-- -->(<!-- -->3<!-- -->)</button></div></div><div class="groups_ums2"><section class="group_Hafd"><h2 class="groupTitle_vOxn">Research Publications</h2><div class="list_fpfs"><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">vLLM Semantic Router: Signal Driven Decision Routing for Mixture-of-Modality Models</h3><p class="rowMeta_x97x">vLLM Semantic Router Team<!-- --> · <!-- -->arXiv Technical Report</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2603.04444" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">The Workload-Router-Pool Architecture for LLM Inference Optimization: A Vision Paper from the vLLM Semantic Router Project</h3><p class="rowMeta_x97x">
27Huamin Chen, Xunzhuo Liu, Bowei He, Fuyuan Lyu, Yankai Chen, Xue Liu, Yuhan Liu, Junchen Jiang<!-- --> · <!-- -->arXiv Technical Report</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2603.21354" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">Visual Confused Deputy: Exploiting and Defending Perception Failures in Computer-Using Agents</h3><p class="rowMeta_x97x">Xunzhuo Liu, Bowei He, Xue Liu, Andy Luo, Haichen Zhang, Huamin Chen<!-- --> · <!-- -->arXiv Technical Report</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2603.14707" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">Outcome-Aware Tool Selection for Semantic Routers: Latency-Constrained Learning Without LLM Inference</h3><p class="rowMeta_x97x">Huamin Chen, Xunzhuo Liu, Junchen Jiang, Bowei He, Xue Liu<!-- --> · <!-- -->arXiv Technical Report</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2603.13426" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">Adaptive Vision-Language Model Routing for Computer Use Agents</h3><p class="rowMeta_x97x">Xunzhuo Liu, Bowei He, Xue Liu, Andy Luo, Haichen Zhang, Huamin Chen<!-- --> · <!-- -->arXiv Technical Report</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2603.12823" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">98× Faster LLM Routing Without a Dedicated GPU: Flash Attention, Prompt Compression, and Near-Streaming for the vLLM Semantic Router</h3><p class="rowMeta_x97x">Xunzhuo Liu, Bowei He, Xue Liu, Andy Luo, Haichen Zhang, Huamin Chen<!-- --> · <!-- -->arXiv Technical Report</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2603.12646" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">inference-fleet-sim: A Queueing-Theory-Grounded Fleet Capacity Planner for LLM Inference</h3><p class="rowMeta_x97x">Huamin Chen, Xunzhuo Liu, Yuhan Liu, Junchen Jiang, Bowei He, Xue Liu<!-- --> · <!-- -->arXiv Technical Report</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2603.16054" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">FleetOpt: Analytical Fleet Provisioning for LLM Inference with Compress-and-Route as Implementation Mechanism</h3><p class="rowMeta_x97x">Huamin Chen, Xunzhuo Liu, Yuhan Liu, Junchen Jiang, Bowei He, Xue Liu<!-- --> · <!-- -->arXiv Technical Report</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2603.16514" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">The 1/W Law: An Analytical Study of Context-Length Routing Topology and GPU Generation Gains for LLM Inference Energy Efficiency</h3><p class="rowMeta_x97x">Huamin Chen, Xunzhuo Liu, Yuhan Liu, Junchen Jiang, Bowei He, Xue Liu<!-- --> · <!-- -->arXiv Technical Report</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2603.17280" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">Beyond Document Grounding: Span-Level Hallucination Detection over Code, Tool Output, and Documents</h3><p class="rowMeta_x97x">Ádám Kovács, Bowei He, Xue Liu, István Boros, Szilveszter Tóth, Gábor Recski<!-- --> · <!-- -->EMNLP 2026 - GroundLM Workshop</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2607.00895" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a><a href="https://groundlm.github.io/grouplm_emnlp2026/" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Event page<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">Conflict-Free Policy Languages for Probabilistic ML Predicates: A Framework and Case Study with the Semantic Router DSL</h3><p class="rowMeta_x97x">Xunzhuo Liu, Hao Wu, Huamin Chen, Bowei He, Xue Liu<!-- --> · <!-- -->arXiv Technical Report</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2603.18174" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">From Inference Routing to Agent Orchestration: Declarative Policy Compilation with Cross-Layer Verification</h3><p class="rowMeta_x97x">
27Huamin Chen, Xunzhuo Liu, Bowei He, Xue Liu<!-- --> · <!-- -->arXiv Technical Report</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2603.27299" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">Knowledge Access Beats Model Size: Memory Augmented Routing for Persistent AI Agents</h3><p class="rowMeta_x97x">Xunzhuo Liu, Bowei He, Xue Liu, Andy Luo, Haichen Zhang, Huamin Chen<!-- --> · <!-- -->arXiv Technical Report</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2603.23013" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">Fast and Faithful: Real-Time Verification for Long-Document Retrieval-Augmented Generation Systems</h3><p class="rowMeta_x97x">Xunzhuo Liu, Bowei He, Xue Liu, Haichen Zhang, Huamin Chen<!-- --> · <!-- -->SIGIR 2026 Industry Track</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2603.23508" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2026</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">Token-Budget-Aware Pool Routing for Cost-Efficient LLM Inference</h3><p class="rowMeta_x97x">Huamin Chen, Xunzhuo Liu, Junchen Jiang, Bowei He, Xue Liu<!-- --> · <!-- -->arXiv Technical Report</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2604.09613" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2025</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">When to Reason: Semantic Router for vLLM</h3><p class="rowMeta_x97x">Chen Wang, Xunzhuo Liu, Yuhan Liu, Yue Zhu, Xiangxi Mo, Junchen Jiang, Huamin Chen<!-- --> · <!-- -->NeurIPS - MLForSys</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2510.08731" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2025</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">Category-Aware Semantic Caching for Heterogeneous LLM Workloads</h3><p class="rowMeta_x97x">Chen Wang, Xunzhuo Liu, Yue Zhu, Alaa Youssef, Priya Nagpurkar, Huamin Chen</p></div><div class="rowLinks_OE72"><a href="https://arxiv.org/abs/2510.26835" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2025</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">Semantic Inference Routing Protocol (SIRP)</h3><p class="rowMeta_x97x">Huamin Chen, Luay Jalil<!-- --> · <!-- -->Internet Engineering Task Force (IETF)</p></div><div class="rowLinks_OE72"><a href="https://datatracker.ietf.org/doc/html/draft-chen-nmrg-semantic-inference-routing" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2025</span><span class="rowType_WH8k rowTypePaper_zA8q">Paper</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">Multi-Provider Extensions for Agentic AI Inference APIs</h3><p class="rowMeta_x97x">H. Chen, L. Jalil, N. Cocker<!-- --> · <!-- -->Internet Engineering Task Force (IETF) - Network Management Research Group</p></div><div class="rowLinks_OE72"><a href="https://www.ietf.org/archive/id/draft-chen-nmrg-multi-provider-inference-api-00.html" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Paper<span aria-hidden="true">↗</span></a></div></article></div></section><section class="group_Hafd"><h2 class="groupTitle_vOxn">Conference Presentations</h2><div class="list_fpfs"><article class="row__Tzd"><span class="rowYear_ruHw">2025</span><span class="rowType_WH8k rowTypeTalk_eL3O">Talk</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">Intelligent LLM Routing: A New Paradigm for Multi-Model AI Orchestration in Kubernetes</h3><p class="rowMeta_x97x">Chen Wang, Huamin Chen<!-- --> · <!-- -->KubeCon NA 2025</p></div><div class="rowLinks_OE72"><a href="https://kccncna2025.sched.com/event/27FaI?iframe=no" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Event page<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2025</span><span class="rowType_WH8k rowTypeTalk_eL3O">Talk</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">vLLM Semantic Router: Unlock the Power of Intelligent Routing</h3><p class="rowMeta_x97x">Xunzhuo Liu<!-- --> · <!-- -->vLLM Meetup Beijing</p></div><div class="rowLinks_OE72"><a href="https://drive.google.com/drive/folders/1nQJ8ZkLSjKxvu36sSHaceVXtttbLvvu-" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">
27Watch recording<span aria-hidden="true">↗</span></a></div></article><article class="row__Tzd"><span class="rowYear_ruHw">2025</span><span class="rowType_WH8k rowTypeTalk_eL3O">Talk</span><div class="rowBody_YM8Z"><h3 class="rowTitle_Jiui">AI-Powered vLLM Semantic Router</h3><p class="rowMeta_x97x">Huamin Chen<!-- --> · <!-- -->vLLM Office Hours</p></div><div class="rowLinks_OE72"><a href="https://www.youtube.com/live/b-ciRqvbtsk" class="rowLink_OnCl" target="_blank" rel="noopener noreferrer">Watch recording<span aria-hidden="true">↗</span></a></div></article></div></section></div></article></div></main></div></div><footer class="theme-layout-footer footer footer--dark"><div class="container container-fluid"><div class="row footer__links"><div class="theme-layout-footer-column col footer__col"><div class="footer__title">Documentation</div><ul class="footer__items clean-list"><li class="footer__item"><a class="footer__link-item" href="/docs/intro">Introduction</a></li><li class="footer__item"><a class="footer__link-item" href="/docs/installation">Quick Start</a></li><li class="footer__item"><a class="footer__link-item" href="/community/governance">Governance</a></li><li class="footer__item"><a class="footer__link-item" href="/community/contributing">Contributing</a></li></ul></div><div class="theme-layout-footer-column col footer__col"><div class="footer__title">Community</div><ul class="footer__items clean-list"><li class="footer__item"><a href="https://github.com/vllm-project/semantic-router" target="_blank" rel="noopener noreferrer" class="footer__link-item">GitHub<svg width="13.5" height="13.5" aria-label="(opens in new tab)" class="iconExternalLink_nPIU"><use href="#theme-svg-external-link"></use></svg></a></li><li class="footer__item"><a href="https://huggingface.co/vllm-sr" target="_blank" rel="noopener noreferrer" class="footer__link-item">Hugging Face<svg width="13.5" height="13.5" aria-label="(opens in new tab)" class="iconExternalLink_nPIU"><use href="#theme-svg-external-link"></use></svg></a></li><li class="footer__item"><a href="https://github.com/vllm-project/semantic-router/discussions" target="_blank" rel="noopener noreferrer" class="footer__link-item">GitHub Discussions<svg width="13.5" height="13.5" aria-label="(opens in new tab)" class="iconExternalLink_nPIU"><use href="#theme-svg-external-link"></use></svg></a></li><li class="footer__item"><a class="footer__link-item" href="/community/contributors">Leaderboard</a></li></ul></div><div class="theme-layout-footer-column col footer__col"><div class="footer__title">More</div><ul class="footer__items clean-list"><li class="footer__item"><a class="footer__link-item" href="/blog">Blog</a></li><li class="footer__item"><a class="footer__link-item" href="/publications">Publications</a></li><li class="footer__item"><a class="footer__link-item" href="/white-paper">White Paper</a></li><li class="footer__item"><a class="footer__link-item" href="/vision-paper">Vision Paper</a></li><li class="footer__item"><a class="footer__link-item" href="/decision-paper">Decision Paper</a></li><li class="footer__item"><a href="https://github.com/vllm-project/semantic-router/blob/main/LICENSE" target="_blank" rel="noopener noreferrer" class="footer__link-item">License<svg width="13.5" height="13.5" aria-label="(opens in new tab)" class="iconExternalLink_nPIU"><use href="#theme-svg-external-link"></use></svg></a></li></ul></div></div><div class="footer__bottom text--center"><div class="footer__copyright">Copyright © 2026 vLLM Semantic Router Team.</div></div></div></footer></div></div>
28
29<script async id="netlify-rum-container" src="/.netlify/scripts/rum" data-netlify-rum-site-id="cead41ef-ca9e-4b52-afbb-a287dde3aa1c" data-netlify-deploy-branch="main" data-netlify-deploy-context="production" data-netlify-cwv-token="eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzaXRlX2lkIjoiY2VhZDQxZWYtY2E5ZS00YjUyLWFmYmItYTI4N2RkZTNhYTFjIiwiYWNjb3VudF9pZCI6IjY4YmI0ODRkNDU2ZGQ3MjZkYzI3MzdhMSIsImRlcGxveV9pZCI6IjZhYzE1OTY1NWM2YTA2MDAwODdkMGQ2OCIsImlzc3VlciI6Im5mc2VydmVyIn0.n2ePjH0F2A6zIbEDG5TqFVXWjdREQ_HoI1Ca8w47Kn0"></script>
29
30</body>
31</html>

Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.