AI conference papers

AI conference paper library

Search papers from 30+ top conferences, including NeurIPS, ICLR, ICML, CVPR, ACL and AAAI. Filter by keyword, then open any paper for its abstract, authors and links.

arXiv

PEMAND: Persona-Enriched Multi-Agent Negotiation for Household Decision-Making

Yuran Sun, Mustafa Sameen, Yaotian Zhang, Rongguan Gu, Mrunal Vibhute, Chia-yu Wu, Yuanyuan Lei, Xilei Zhao

Modeling household-level decisions is central to many real-world applications, including trip planning, residential mobility and migration, disaster management, etc. Existing studies primarily rely on classical machine learning models with limited predictive capacity, while recent LLM-based approaches have yet to incorporate behavioral theory or intra-household interaction dynamics, both of which are essential for modeling realistic household decisions. To address these limitations, we propose Persona-Enriched Multi-Agent Negotiation for household Decision-making (PEMAND), a novel LLM-based framework that integrates behavioral theory into individualized, household-aware persona modeling and simulates household-level decision-making through structured multi-agent negotiation. Specifically, PEMAND transforms static sociodemographic attributes into coherent narrative profiles that explicitly encode household-level attitudes, subjective norms, and perceived behavioral controls, following our proposed Household-Aware Chain-of-Planned-Behavior (HA-CoPB) framework. Building on these theory-grounded personas, PEMAND captures real-world household decision negotiation via a structured two-phase multi-agent conversation framework with a novel persona-alignment control mechanism. Evaluated on national and regional household decision datasets across two major domains, including travel behavior and residential mobility, PEMAND consistently outperforms state-of-the-art benchmarks.

arXiv

Compiled AI: Deterministic Code Generation for LLM-Based Workflow Automation

Geert Trooskens, Aaron Karlsberg, Anmol Sharma, Lamara De Brouwer, Max Van Puyvelde, Matthew Young, John Thickstun, Gil Alterovitz, Walter A. De Brouwer

We study compiled AI, a paradigm in which large language models generate executable code artifacts during a compilation phase, after which workflows execute deterministically without further model invocation. This paradigm has antecedents in prior work on declarative pipeline optimization (DSPy) and hybrid neural-symbolic planning (LLM+P); our contribution is a systems-oriented study of its application to high-stakes enterprise workflows, with particular emphasis on healthcare settings where reliability and auditability are critical. By constraining generation to narrow business-logic functions embedded in validated templates, compiled AI trades runtime flexibility for predictability, auditability, cost efficiency, and reduced security exposure. We introduce (i) a system architecture for constrained LLM-based code generation, (ii) a four-stage generation-and-validation pipeline that converts probabilistic model output into production-ready code artifacts, and (iii) an evaluation framework measuring operational metrics including token amortization, determinism, reliability, security, and cost. We evaluate on two task types: function-calling (BFCL, n=400) and document intelligence (DocILE, n=5,680 invoices). On function-calling, compiled AI achieves 96% task completion with zero execution tokens, breaking even with runtime inference at approximately 17 transactions and reducing token consumption by 57x at 1,000 transactions. On document intelligence, our Code Factory variant matches Direct LLM on key field extraction (KILE: 80.0%) while achieving the highest line item recognition accuracy (LIR: 80.4%). Security evaluation across 135 test cases demonstrates 96.7% accuracy on prompt injection detection and 87.5% on static code safety analysis with zero false positives.

arXiv

Generative AI in Action: Field Experimental Evidence from Alibaba's Customer Service Operations

Xiao Ni, Yiwei Wang, Tianjun Feng, Lauren Xiaoyan Lu, Yitong Wang, Congyi Zhou

In collaboration with Alibaba, we study how a generative AI assistant affects service performance in e-commerce after-sales operations. In a large-scale field experiment, human agents providing digital chat support were randomly assigned access to a gen AI assistant. The assistant drafts issue diagnoses and solution proposals in the opening stage only; agents can adopt, modify, or disregard them. Because of this discretion, we estimate the effects of both gen AI access and usage. On average, gen AI improves service speed and subjective service quality, measured by customer ratings, but has no significant effect on objective service quality, measured by customer retrials. These gains come from more than automation. Gen AI reshapes agent-customer interactions: treated agents respond faster and take a more proactive role, while customers provide less input; both patterns persist into later chat stages. These average effects, however, mask heterogeneity across agents. Lower-performing agents benefit the most, indicating that gen AI can narrow performance gaps. Top-performing agents experience declines in both subjective and objective service quality. This decline is consistent with workflow disruption: among top performers, gen AI use increases shift-away time, response delays, and immediate customer retrials, suggesting weakened service continuity in the focal chat. Successful gen AI deployment therefore requires careful performance evaluation and rollout tailored to agent skill.

arXiv

Shaping Scientific Explanations to Expert Perspectives with Persona-Conditioned Reinforcement Learning

Susana Nunes, Tiago Guerreiro, Catia Pesquita

Explainable AI is increasingly important to scientific discovery. However, existing methods largely ignore that explanation quality is not universal: experts differ in how they assess evidence, prioritize mechanisms, and construct explanatory narratives. We introduce perspective-conditioned explanations, a framework for adapting explanation generation to epistemic variation in expert judgment. Using knowledge graph reasoning paths in drug discovery, we show that preferences organize into coherent epistemic perspectives that can be captured by agentic personas, representations of how experts evaluate explanations. Persona-aligned rewards then guide reinforcement learning-based explanation generation without large-scale expert supervision. Expert user studies show that perspective-conditioned explanations are preferred over general-purpose explanations and improve perceived relevance and validity. Moreover, they match or exceed state-of-the-art predictive performance and reduce expert feedback time by two orders of magnitude. Together, these findings demonstrate that explanation quality is perspective-dependent and that modeling this variation enables scalable and human-aligned explanation generation for scientific discovery.

arXiv

Agentic Harness for Real-World Compilers

Yingwei Zheng, Cong Li, Shaohua Li, Yuqun Zhang, Zhendong Su

Compilers are critical to modern computing, yet fixing compiler bugs is difficult. While recent large language model (LLM) advancements enable automated bug repair, compiler bugs pose unique challenges due to their complexity, deep cross-domain expertise requirements, and sparse, non-descriptive bug reports, necessitating compiler-specific harnesses. To bridge the gap, we introduce llvm-harness, the first harness designed to assist LLM agents in understanding and fixing compiler bugs. Our current focus is on the middle end of LLVM, one of the most widely used compiler infrastructures. Central to llvm-harness are agent-friendly LLVM tools, a benchmark llvm-bench of 334 reproducible LLVM middle-end bugs, and a tailored mini agent llvm-autofix-mini for fixing LLVM middle-end bugs automatically. We evaluate five frontier models and find that they exhibit a performance decline when tackling compiler bugs with the state-of-the-art agent. With llvm-harness' enhancement, their performance improves by 62%. Our specialized mini agent llvm-autofix-mini further outperforms the llvm-harness-enhanced state-of-the-art by 22%. This emphasizes the necessity for specialized harnesses like ours to assist LLMs in compiler engineering tasks. Despite promising results, our expert review also reveals several open challenges that remain when applying LLMs for compiler engineering tasks. GitHub: https://github.com/dtcxzyw/llvm-harness

arXiv

ELISA: An Interpretable Hybrid Generative AI Agent for Expression-Grounded Discovery in Single-Cell Genomics

Omar Coser

Translating single-cell RNA sequencing (scRNA-seq) data into mechanistic biological hypotheses remains a critical bottleneck, as agentic AI systems lack direct access to transcriptomic representations while expression foundation models remain opaque to natural language. Here we introduce ELISA (Embedding-Linked Interactive Single-cell Agent), an interpretable framework that unifies scGPT expression embeddings with BioBERT-based semantic retrieval and LLM-mediated interpretation for interactive single-cell discovery. An automatic query classifier routes inputs to gene marker scoring, semantic matching, or reciprocal rank fusion pipelines depending on whether the query is a gene signature, natural language concept, or mixture of both. Integrated analytical modules perform pathway activity scoringacross 60+ gene sets, ligand-receptor interaction prediction using 280+ curated pairs, condition-aware comparative analysis, and cell-type proportion estimation all operating directly on embedded data without access to the original count matrix. Benchmarked across six diverse scRNA-seq datasets spanning inflammatory lung disease, pediatric and adult cancers, organoid models, healthy tissue, and neurodevelopment, ELISA significantly outperforms CellWhisperer, a classical lexical retriever (BM25), and a random baseline in cell type retrieval (combined permutation test, $p < 2\times10^{-5}$ for each), with particularly large gains on gene-signature queries (Cohen's $d = 5.98$ for MRR). ELISA replicates published biological findings (mean composite score 0.88), and generates candidate hypotheses through grounded LLM reasoning, bridging the gap between transcriptomic data exploration and biological discovery.

arXiv

Differentially Private Distributed Inference for Multicenter Clinical Studies

Marios Papachristou, M. Amin Rahimian

Extracting reliable conclusions from data distributed across institutions is a core problem in healthcare: pooling patient records would improve inference, but privacy regulations and the lack of a trusted central authority frequently delay multicenter studies. We develop a framework for differentially private distributed inference in which institutions repeatedly exchange log belief-ratio statistics subject to differential privacy (DP). With arithmetic and geometric averaging of beliefs, we control the false-negative and false-positive rates as functions of the privacy budget, communication rounds, and statistical separation between hypotheses, exposing a three-way trade-off among accuracy, communication, and privacy. We derive finite-sample bounds on the Type I and Type II error probabilities that allow distributed hypothesis testing at a target significance level, and show that the Laplace mechanism minimizes convergence time subject to DP. For distributed online learning from data streams (e.g., epidemiological surveillance or rolling recruitment), privacy noise vanishes asymptotically and online learning admits similar finite-sample guarantees. On simulated multicenter survival analyses using the AIDS Clinical Trials Group and an advanced-cancer cohort, and a simulated genetic association study over New York City hospitals, our method approaches the non-private baseline at a small privacy budget between 1 and 10, runs 10x to 1000x faster than homomorphic-encryption methods, and incurs up to 100x lower error than first-order private optimization methods. Finally, the level of aggregation is a primary design choice: federating at the organizational rather than the hospital level strengthens privacy, raises statistical power, and lowers communication and administrative burden, so data should be pooled within organizations before setting up federated analytics.

arXiv

AIvilization v0: Toward Large-Scale Artificial Social Simulation with a Unified Agent Architecture and Adaptive Agent Profiles

Wenkai Fan, Shurui Zhang, Xiaolong Wang, Haowei Yang, Tsz Wai Chan, Xingyan Chen, Junquan Bi, Zirui Zhou, Jia Liu, Kani Chen

AIvilization v0 is a publicly deployed large-scale artificial society that couples a resource-constrained sandbox with a unified LLM-agent architecture, aiming to sustain long-horizon autonomy while remaining executable under a rapidly changing environment. To mitigate the tension between goal stability and reactive correctness, keeping long-horizon objectives on course while each action remains valid in a fast-changing shared world, we introduce (i) a hierarchical branch-thinking planner that decomposes life goals into parallel objective branches and uses simulation-guided validation plus tiered re-planning to ensure feasibility; (ii) an adaptive agent profile with dual-process memory that separates short-term execution traces from long-term semantic consolidation, enabling persistent yet evolving identity; and (iii) a human-in-the-loop steering interface that injects long-horizon objectives and short commands at appropriate abstraction levels, with effects propagated through memory instead of brittle prompt overrides. The environment integrates physiological survival costs, non-substitutable multi-tier production, an AMM-based price mechanism, and a gated education-occupation system. In a large-scale public deployment with tens of thousands of agents, high-frequency transactions from the platform's mature phase reveal stable markets that reproduce key stylized facts of real economies and structured wealth stratification driven by education and access constraints. At the agent level, portraits evolve coherently over long horizons, and human steering is associated with measurably larger short-horizon profile updates. Controlled ablation experiments complement the deployment evidence, showing that our agent architecture is robust in multi-objective, long-horizon settings.

arXiv

GPU-Accelerated ANNS: Quantized for Speed, Built for Change

Hunter McCoy, Zikun Wang, Prashant Pandey

Approximate nearest neighbor search (ANNS) is a core problem in machine learning and information retrieval applications. GPUs offer a promising path to high-performance ANNS: they provide massive parallelism for distance computations, are readily available, and can co-locate with downstream applications. Despite these advantages, current GPU-accelerated ANNS systems face three key limitations. First, real-world applications operate on evolving datasets that require fast batch updates, yet most GPU indices must be rebuilt from scratch when new data arrives. Second, high-dimensional vectors strain memory bandwidth, but current GPU systems lack efficient quantization techniques that reduce data movement without introducing costly random memory accesses. Third, the data-dependent memory accesses inherent to greedy search make overlapping compute and memory difficult, leading to reduced performance. We present Jasper, a GPU-native ANNS system with both high query throughput and updatability. Jasper builds on the Vamana graph index and overcomes existing bottlenecks via three contributions: (1) a CUDA batch-parallel construction algorithm that enables lock-free streaming insertions, (2) a GPU-efficient implementation of RaBitQ quantization that reduces memory footprint up to 8x without the random access penalties, and (3) an optimized greedy search kernel that increases compute utilization, resulting in better latency hiding and higher throughput. Our evaluation across five datasets shows that Jasper achieves up to 1.84x higher query throughput than CAGRA and achieves up to 80% peak utilization as measured by the roofline model. Jasper's construction scales efficiently and constructs indices an average of 7x faster than CAGRA while providing updatability that CAGRA lacks. Compared to BANG, the previous fastest GPU Vamana implementation, Jasper delivers 10-74x faster queries.

arXiv

Differentially Private Nonparametric Modal Learning with Applications to Regression and Clustering

Arkajyoti Bhattacharjee, Arnab Auddy

Density modes provide a localized and interpretable summary of multimodal distributions, but their estimation under rigorous differential privacy constraints remains largely unexplored. We study differentially private recovery of density modes for multivariate distributions under local smoothness, curvature, and separation conditions. We propose DP-GRAMS, a mean-shift inspired method that performs noisy ascent on a differentially private score estimator. Assuming the density belongs locally to a Hölder class with smoothness parameter $β> 2$, our score estimator uses bias-reducing higher-order kernels, and then enforces privacy in the gradient ascent steps via gradient clipping and calibrated Gaussian noise. A private initialization scheme combines a density-aware utility with a suppression rule and, with $k\asymp M\log n$ draws over a public $h_{\mathrm{DAP}}$-grid and suppression radius $ρ_{\mathrm{init}}\asymp (\log n)^{-1/d}$, achieves high-probability coverage of the modal basins by successively suppressing selected local neighborhoods in competitive regions, while correlated noise across multiple starts enables joint release under a single $(\varepsilon,δ)$-differential privacy guarantee. We prove that all population modes are recovered with high probability and establish asymptotic error rates of the form $O\!\left((\tfrac{\log n}{n})^{\frac{2(β-1)}{d+2β}}\right) + O\!\left((\tfrac{\mathrm{polylog}(n,δ)}{n^2\varepsilon^2})^{\frac{β-1}{d+β}}\right)$. We also provide minimax lower bounds for private mode estimation, and show that our estimators are nearly optimal, up to a logarithmic factor in the MSE. We present two natural extensions: DP-PMS, a private modal-regression method, and DP-GRAMS-C, a clustering pipeline. Extensive experiments on synthetic and real data demonstrate favorable privacy-utility trade-offs relative to common baselines.

arXiv

Sign compression for Muon: SignMuon, MuonSign, and the Limits of Error Feedback

Maria Smirnova, Alexey Kravatskiy

SignMuon compresses the Muon update to one bit per parameter by taking its elementwise sign, providing the most direct way to run a matrix-aware optimizer under an extremely low communication budget. It outperforms SignSGD in practice, yet it can ascend even on a linear function. Signing the gradient before the Linear Minimization Oracle (LMO), rather than after, does not repair this: we construct a small explicit instance on which sign-before (MuonUSign) and sign-on-both-sides (MuonSign) ascend as well, so no placement of the sign around the oracle descends in general. Error feedback, the standard remedy for a biased compressor, does not rescue SignMuon: when applied to Muon's output, error feedback can fail for every smoothness constant, step size, and momentum. Applied to the gradient, error feedback does work, and EF21-MuonUSign and EF21-MuonSign attain the standard $\mathcal{O}(T^{-1/2})$ rate for the squared gradient norm on smooth nonconvex problems, the latter at one bit in each direction. Experiments then reverse the ordering: across centralized CIFAR-10, federated CIFAR-10, and the nanoGPT speedrun, the strongest compressed method is consistently sign-after-the-LMO, precisely the placement we prove divergent, with the provably convergent variants trailing it. Compressing after the LMO, a heuristic, matters more at these scales than the guarantee does.

arXiv

M3MAD-Bench: Multi-Dimensional Evaluation of Multi-Agent Debate Across Domains and Modalities

Ao Li, Jinghui Zhang, Luyu Li, Yuxiang Duan, Lang Gao, MingCai Chen, Weijun Qin, Shaopeng Li, Fengxian Ji, Ning Liu, Lizhen Cui, Xiuying Chen, Yuntao Du

As an agent-level reasoning and coordination paradigm, Multi-Agent Debate (MAD) orchestrates multiple agents through structured debate to improve answer quality and support complex reasoning. However, existing research on MAD suffers from two fundamental limitations: evaluations are conducted under fragmented and inconsistent settings, making fair comparison difficult, and are largely confined to text-only scenarios, leaving its effectiveness in multimodal settings underexplored. To address these gaps, we introduce M3MAD-Bench, a unified and extensible benchmark for evaluating MAD methods across Multi-domain tasks, Multi-modal inputs, and Multi-dimensional metrics. M3MAD-Bench establishes standardized protocols over five core task domains, including Knowledge, Mathematics, Medicine, Natural Sciences, and Complex Reasoning, covering a total of 13 datasets, and systematically includes both pure text and vision-language data, enabling controlled cross-modality comparison. We evaluate MAD methods on 9 base models spanning different architectures, scales, and modality capabilities. Beyond accuracy, M3MAD-Bench incorporates efficiency-oriented metrics such as token consumption and inference time, providing a holistic view of performance--cost trade-offs. Through extensive experiments, we derive nine key insights, revealing that MAD is not uniformly effective: collaborative methods are generally more robust than adversarial ones, especially on reasoning-intensive and multimodal tasks, but often incur substantial efficiency costs. These findings provide practical guidance for selecting and designing MAD strategies in real-world applications. We believe M3MAD-Bench offers a reliable foundation for future research on standardized and reproducible MAD evaluation. The code is available at https://github.com/liaolea/M3MAD-Bench.

arXiv

WebCoderBench: Benchmarking Web Application Generation with Comprehensive and Interpretable Evaluation Metrics

Chenxu Liu, Yingjie Fu, Wei Yang, Ying Zhang, Tao Xie

Web applications (web apps) have become a key arena for large language models (LLMs) to demonstrate their code generation capabilities and commercial potential. However, building a benchmark for LLM-generated web apps remains challenging due to the need for real-world user requirements, generalizable evaluation metrics without relying on ground-truth implementations or test cases, and interpretable evaluation results. To address these challenges, we introduce WebCoderBench, the first real-world-collected, generalizable, and interpretable benchmark for web app generation. WebCoderBench comprises 1,572 real user requirements, covering diverse modalities and expression styles that reflect realistic user intentions. WebCoderBench provides 24 fine-grained evaluation metrics across 9 perspectives, combining rule-based and LLM-as-a-judge paradigm for fully automated, objective, and general evaluation. Moreover, WebCoderBench adopts human-preference-aligned weights over metrics to yield interpretable overall scores. Experiments across 12 representative LLMs and 2 LLM-based agents show that there exists no dominant model across all evaluation metrics, offering an opportunity for LLM developers to optimize their models in a targeted manner for a more powerful version.

arXiv

Freeze, Then Select: Structured Field Adapters and Stability-Validated Weak Selection for PDE Discovery from Sparse Observations

Juncheng Zhong, Chenghuang Shen, Jianfeng Liu, Zhengdong Xiao, Longjiu Luo, Qianrong Wang, Wenjun Xu, Wenlian Lu

PDE discovery from sparse observations requires reconstructing a continuous field and selecting the correct differential terms. Our analysis of optimization paths in coupled neural PDE discovery reveals three behaviors: the exact support can persist to the end of training, appear only transiently, or fail to emerge. To decouple equation selection from neural optimization, we develop a freeze-then-select method combining a structured field adapter with Stability-Validated Weak Selection (SVWS). Trained from observations without a PDE residual, the adapter factorizes the field into learned spatial features and temporal coefficients represented by cubic splines. After freezing the field, SVWS identifies recurrent terms across independent weak-form systems, refits candidate supports, and selects the final equation on held-out weak-form systems. Beyond fixed libraries, we apply the same principle to expressions generated by genetic programming and recover the power-law form of an unknown nonlinear diffusion function from sparse, noisy observations. Across all six sparse MDBench regimes, our method attains the highest exact support recovery rate, with its clearest gains over classical and neural baselines on challenging Kuramoto-Sivashinsky dynamics.

arXiv

CENDRe: Concept Extraction with Natural Domain Representations

Antonia Holzapfel, Andres Felipe Posada-Moreno, Sebastian Trimpe

Convolutional neural networks (CNNs) are widely used for time-series classification, but their deployment in critical domains requires understanding the temporal and spectral patterns that drive their predictions. Concept extraction (CE) methods identify such patterns by analyzing representations within the models' latent space. However, existing time-series CE methods have three limitations: they operate only in the time domain and overlook frequency features, predefine the number of concepts, and produce localizations misaligned with the regions the model uses. We address these limitations by proposing CENDRe, a concept extraction method for CNNs. It first discovers concepts by clustering per-timestep latent representations in two stages, where silhouette-guided aggregation selects the number of concepts automatically. Then, it localizes each concept through gradients of a presence score that contrasts the latent representations with their prototypes, producing masks that concentrate on the regions driving the concept. These gradients, propagated through a differentiable invertible mapping of the input such as a Fourier transform, yield localizations for the same concepts in the frequency domain. Finally, each concept receives a relevance score that quantifies its contribution to each class. On synthetic benchmarks, CENDRe achieves representation correctness comparable to state-of-the-art CE methods and significantly higher importance correctness. On real bearing-fault data, CENDRe extracts the frequency bands driving the model's predictions, located in regions commonly inspected for fault diagnosis, producing evidence to assess the model that time-domain CE methods cannot.

arXiv

When Does On-Policy Interaction Help? Representational Tradeoffs in Value-Based Imitation Learning

Luca Viano, Antoine Moulin, Audrey Huang, Volkan Cevher, Philip Amortila, Dylan J Foster

Imitation learning (IL)---training an agent to replicate expert behavior from demonstrations---underpins applications from robotics to language model training. Standard approaches such as Behavior Cloning (BC) are known to suffer from compounding errors and performance plateaus, particularly when the learner cannot perfectly represent the expert's policy (as is typical, e.g., in distillation). Two interventions are widely understood empirically to improve performance: querying the expert interactively along the learner's own trajectories, and using value function estimation en route to generating a policy rather than directly fitting the expert's full action distribution. We investigate the nature of these improvements and their potentially surprising interplay. Our main finding is that expert interaction relaxes the representational demands on the learner: one only needs a model capable of realizing the expert's value function, bypassing the (often stricter) requirement of realizing the expert's policy itself. Concretely, we introduce OVI, an interactive on-policy IL algorithm that is statistically efficient whenever the learner can represent the expert's value function and computationally efficient given access to a linear maximization oracle. We complement this with a negative result showing that interaction is necessary. Namely, without stronger assumptions beyond expert-value realizability alone, any offline IL algorithm must scale with the complexity of the expert policy class. Our findings bear out empirically. OVI outperforms offline policy-based (BC), interactive policy-based (DAgger), and offline value-based IL methods, with the largest gains when the learner network is substantially less expressive than the expert's.

arXiv

Pay for The Second-Best Service: A Game-Theoretic Approach Against Dishonest LLM Providers

Yuhan Cao, Yu Wang, Sitong Liu, Miao Li, Yixin Tao, Tianxing He

The widespread adoption of Large Language Models (LLMs) through Application Programming Interfaces (APIs) induces a critical vulnerability: the potential for dishonest manipulation by service providers. This manipulation can manifest in various forms, such as secretly substituting a proclaimed high-performance model with a low-cost alternative, or inflating responses with meaningless tokens to increase billing. This work tackles the issue through the lens of algorithmic game theory and mechanism design. We are the first to propose a formal economic model for a realistic user-provider ecosystem, where a user can iteratively delegate $T$ queries to multiple model providers, and providers can engage in a range of strategic behaviors. As our central contribution, we prove that for a continuous strategy space and any $ε\in(0,\frac12)$, there exists an approximate incentive-compatible mechanism with an additive approximation ratio of $O(T^{1-ε}\log T)$, and a guaranteed quasi-linear second-best user utility. We also prove an impossibility result, stating that no mechanism can guarantee an expected user utility that is asymptotically better than our mechanism. Furthermore, we demonstrate the effectiveness of our mechanism in simulation experiments with real-world API settings.

arXiv

Shall We Play a Game? Language Models for Open-ended Wargames

Glenn Matlin, Isaac Song, Yixiong Hao, Parv Mahajan, Evan Montoya, Ryan Bard, Stuart R. Topp, Anthony Wen-Ming Zang, Mohammed Rehan Parwani, Soham Shetty, Mark Riedl

LLM-based social simulations can make a generated transcript look like a single behavioral signal, but the model behind that transcript may be doing several different jobs: choosing what an actor says or does, deciding what happens after an action, or both. The difference matters especially in open-ended wargames, where models are prized for handling unusual actions and ambiguous consequences. We report a scoping review of 223 de-duplicated AI-in-wargames and strategic-simulation papers retrieved through May 1, 2026, describing each simulation by its model-control profile: whether the language model has open-ended control over player actions, adjudication, or both. Only 20 of 223 studies (~9%) give language models both roles. Before treating LM outputs as social simulations, researchers need to know how much creative control the model has over actions and consequences. For open-ended simulations, fidelity depends not only on whether agents behave plausibly, but also on whether language models can reliably act as adjudicators or world models.

arXiv

QASP: Query-Adaptive Robust Vector Search Policy

Hakan Ferhatosmanoglu, Kushal Kumar, Tal Wagner, Andy Warfield

A fundamental challenge of vector search is achieving consistently high recall while minimizing computational costs. Fixed search parameters cause significant performance variance across queries, and conventional evaluation on average recall masks these per-query disparities. We introduce QASP (Query-Adaptive robust vector Search Policy), which predicts the complete recall progression curve per query via a single upfront supervised regression, from which a search policy is derived for any recall target; this avoids iterative model invocations during search or separate predictors per target. By predicting normalized recall values with scale-invariant features and pre-search inference, QASP generalizes across recall targets, index configurations, and datasets. Its fine-grained progress predictions further enable a lightweight reactive complement that adjusts search depth based on predicted-versus-observed deviations without additional inference. We prove that QASP requires a finite training sample independent of dataset size and dimensionality, that its loss exceeds the irreducible lower bound of any fixed policy by a vanishing margin, and that its data access savings over fixed probing grow exponentially in intrinsic dimensionality. Experimentally, QASP achieves significantly lower recall variance and deviation from target, higher query satisfaction rate, and scales to large data and hierarchical indices without retraining, achieving 99% recall with 80% less data access.

arXiv

The Parts Are Greater Than the Sum: Automated Task Sequencing for Efficient Training of Multi-Policy LLMs

Jiajia Tang, Sizhe Yuen, Francisco Gomez Medina, Yali Du, Adam Sobey

Parameter-Efficient Fine-Tuning (PEFT) commonly adapts large language models using a single shared Low-Rank Adapter (LoRA). This shared optimization space often suffers from interference when adapting heterogeneous task sequences, leading to poor transfer and catastrophic forgetting. Existing approaches mainly improve adapter expressiveness by increasing parameter capacity or composing multiple adapters, yet they still rely on a shared optimization path. In this paper, we propose an optimization-path organization framework for parameter-efficient fine-tuning of large language models, implemented as an automatic multi-policy PEFT architecture. Specifically, optimization-compatible adaptation paths are automatically organized through task grouping and task sequencing under a fixed parameter budget. The organized optimization paths are implemented as independent Quantized Low-Rank Adapters (QLoRA), enabling heterogeneous tasks to be optimized in decoupled adaptation spaces while preserving positive transfer among compatible tasks. Experiments on the TRACE benchmark demonstrate that performance consistently improves from conventional single-policy PEFT to multi-policy PEFT, with the proposed automatic multi-policy framework achieving the best performance of 44.78 under the same trainable capacity. This suggests that optimization-path organization is more effective than simply increasing adapter capacity for heterogeneous parameter-efficient fine-tuning.

Keep browsing

More papers from top AI conferences

30+ top conferences, by conference, year, and topic.