Singapore

Offline multi-agent reinforcement learning (MARL) is severely hampered by the challenge of evaluating out-of-distribution (OOD) joint actions. Our core finding is that when the behavior policy is factorized—a common scenario where agents act fully or partially independently during data collection—a strategy of partial action replacement (PAR) can significantly mitigate this challenge. PAR updates a single or part of agents&#39; actions while the others remain fixed to the behavioral data, reducing distribution shift compared to full joint-action updates. Based on this insight, we develop Soft-Partial Conservative Q-Learning (SPaCQL), using PAR to mitigate OOD issue and dynamically weighting different PAR strategies based on the uncertainty of value estimation. We provide a rigorous theoretical foundation for this approach, proving that under factorized behavior policies, the induced distribution shift scales linearly with the number of deviating agents rather than exponentially with the joint-action space. This yields a provably tighter value error bound for this important class of offline MARL problems. Our theoretical results also indicate that SPaCQL adaptively addresses distribution shift using uncertainty-informed weights. Our empirical results demonstrate SPaCQL enables more effective policy learning, and manifest its remarkable superiority over baseline algorithms when the offline dataset exhibits the independence structure.

AAAI 2026

Partial Action Replacement: Tackling Distribution Shift in Offline MARL

Offline multi-agent reinforcement learning (MARL) is severely hampered by the challenge of evaluating out-of-distribution (OOD) joint actions. Our core finding is that when the behavior policy is factorized—a common scenario where agents act fully or partially independently during data collection—a strategy of partial action replacement (PAR) can significantly mitigate this challenge. PAR updates a single or part of agents' actions while the others remain fixed to the behavioral data, reducing distribution shift compared to full joint-action updates. Based on this insight, we develop Soft-Partial Conservative Q-Learning (SPaCQL), using PAR to mitigate OOD issue and dynamically weighting different PAR strategies based on the uncertainty of value estimation. We provide a rigorous theoretical foundation for this approach, proving that under factorized behavior policies, the induced distribution shift scales linearly with the number of deviating agents rather than exponentially with the joint-action space. This yields a provably tighter value error bound for this important class of offline MARL problems. Our theoretical results also indicate that SPaCQL adaptively addresses distribution shift using uncertainty-informed weights. Our empirical results demonstrate SPaCQL enables more effective policy learning, and manifest its remarkable superiority over baseline algorithms when the offline dataset exhibits the independence structure.

poster

We are pleased to announce the Fortieth AAAI Conference on Artificial Intelligence (AAAI-26), which will be held in Singapore EXPO from January 20 to January 27, 2026.

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. AAAI-26 will feature technical paper presentations, special tracks, invited speakers, workshops, tutorials, poster sessions, senior member presentations, competitions, and exhibit programs, and a range of other activities to be announced.<br><br>

To access this event page, you need to log in with the **email address you registered with**. <br>Access credentials will be sent to your email from Underline -  subject line "Welcome to AAAI 2026". Please be sure to check your spam email folder if you do not see an email confirmation right away.

Please log in

To access this event page, you are required to register.
Please complete your registration to continue.

We recommend reading [**the registration information**](https://aaai.org/conference/aaai/aaai-26/registration/) first.

**Online Registration Form**: https://aaai.getregistered.net/conference-2026 

Registration Required

We are pleased to announce the Fortieth AAAI Conference on Artificial Intelligence (AAAI-26), which will be held in Singapore EXPO from January 20 to January 27, 2026.

We introduce a new task of open-world object counting in videos: given a text description, or an image example, that specifies the target object, the objective is to enumerate all the unique instances of the target objects in the video. This task is especially challenging in crowded scenes with occlusions and objects of similar appearance, where avoiding double counting and identifying reappearances is crucial. To this end, we make the following contributions: we introduce a model, CountVid, for this task. It leverages an image-based counting model, and a promptable video segmentation and tracking model, to enable automated open-world object counting across video frames. To evaluate its performance, we introduce VideoCount, a new dataset for this novel task built from the TAO and MOT20 tracking datasets, as well as from videos of penguins and metal alloy crystallization captured by x-rays. Using this dataset, we demonstrate that CountVid provides accurate object counts, and significantly outperforms strong baselines. The VideoCount dataset, the CountVid model, and all the code will be publicly released.

Open-World Object Counting in Videos

Multi-unit bilateral trade refers to the setting, where there is a buyer and a seller, who holds a finite number of units of an indivisible item. An automated mechanism has to decide how many units are transferred from the seller to the buyer and the corresponding payment from the buyer to the seller. The buyer and the seller have both either increasing or increasing submodular valuation functions in the number of units in possession. The (single-unit) bilateral trade problem arises as a particular case.

We study the problem of social welfare maximisation by establishing the fraction (\emph{approximation ratio}) of the optimal social welfare that a fixed-price mechanism can recover. Fixed-price mechanisms, understood as per-unit price in the multi-unit setting, have been characterised as the only truthful, individually rational and strongly budget balanced mechanisms by (Gerstgrasser et al. 2019) and (Hagerty and Rogerson 1987). We narrow the gap on the approximation ratio of optimal fixed-price mechanisms for bilateral trade, which has been shown to lie between $0.72$ and $0.7381$ by (Cai and Wu 2023). We show that it must lie between $0.728$ and $0.73805$, which leads to improved bounds on the approximation ratio of optimal fixed-price mechanisms for multi-unit bilateral trade. In particular, we show that multi-unit bilateral trade is at least as hard as single-unit bilateral trade, and obtain several hardnesses for different numbers of units.

On the Approximation Ratio of Optimal Fixed-Price Mechanisms for Single and Multi-Unit Bilateral Trade

Probabilistic decoding in Large Language Models (LLMs) often yields inconsistent outputs, particularly on complex or long-form questions. Self-Consistency (SC) mitigates this for short-form QA by majority voting over exact strings, whereas Universal Self-Consistency (USC) and Weighted Unigram Consistency Score (WUCS) extend to long-form responses but lose accuracy on short-form benchmarks. 

We introduce $\textbf{Latent Self-Consistency (LSC)}$, which selects the most semantically consistent response using learnable token embeddings. A lightweight forward generation of summary tokens increases inference time by less than 1% and requires no changes to the model architecture.

Across 6 short-form and 5 long-form reasoning benchmarks (e.g., MATH, MMLU, TruthfulQA), LSC surpasses SC, USC and WUCS on all short-form and long-form ones on average, while maintaining negligible computational overhead. These results position LSC as a practical consistency-selection method that works reliably across answer formats.
Additionally, LSC provides well-calibrated confidence estimates, maintaining low Expected Calibration Error across both answer formats.

Latent Self-Consistency for Reliable Majority-Set Selection in Short- and Long-Answer Reasoning

Large language models (LLMs) have significantly transformed natural language understanding and generation, but they raise privacy concerns due to potential exposure of sensitive information. Studies have highlighted the risk of information leakage, where adversaries can extract sensitive information embedded in the prompts. In this work, we introduce a novel private prediction framework for generating high-quality synthetic text with strong privacy guarantees. Our approach leverages the Differential Privacy (DP) framework to ensure worst-case theoretical bounds on information leakage without requiring any fine-tuning of the underlying models.The proposed method performs inference on private records and aggregates the resulting per-token output distributions. This enables the generation of longer and coherent synthetic text while maintaining privacy guarantees. Additionally, we propose a simple blending operation that combines private and public inference to further enhance utility. Empirical evaluations demonstrate that our approach outperforms previous state-of-the-art methods on in-context-learning (ICL) tasks, making it a promising direction for privacy-preserving text generation while maintaining high utility.

Privacy Preserving In-Context-Learning Framework for Large Language Models

Accurate 3D vehicle pose and shape reconstruction from monocular images remains a formidable challenge for autonomous driving, particularly for distant, occluded, or small objects. Existing methods often suffer from geometric ambiguity in depth estimation and structural hollowness in shape recovery, primarily due to inadequate multi-scale feature aggregation and inflexible prior modeling. To overcome these limitations, a novel framework termed MonoVPR is proposed by integrating dynamic context adaptation and progressive geometry refinement. Specifically, a Hierarchical Dual-Context Attention (HDCA) module is introduced to resolve scale-dependent degradation through gated cross-attention across multi-resolution feature maps, dynamically fusing object-centric geometric cues with scene-centric semantics. For shape refinement, the Bounded Iterative Mesh Refiner (BIMR) is developed, where template-guided deformations are progressively optimized via multi-head deformable attention and a tanh-bounded correction loop, ensuring physically plausible reconstructions. Extensive experiments on the ApolloCar3D benchmark demonstrate MonoVPR achieves state-of-the-art performance, showcasing exceptional capability in reconstructing geometrically consistent shapes and precise poses for challenging long-range and occluded scenarios.

Monocular Vehicle Pose and Shape Reconstruction via Dynamic Context Adaptation and Progressive Geometry Refinement

Automatic presentation slide generation can greatly streamline content creation. However, since preferences of each user may vary, existing under-specified formulations often lead to suboptimal results that fail to align with individual user needs. We introduce a novel task that conditions slide generation on user-specified preferences. We propose a human-behavior-inspired agentic framework, SlideTailor, that progressively generates editable slides in a user-aligned manner. Instead of requiring users to write their preferences in detailed textual form, our system only asks for a paper–slides example pair and a visual template—natural and easy-to-provide artifacts that implicitly encode rich user preferences across content and visual style. Despite the implicit and unlabeled nature of these inputs, our framework effectively distills and generalizes the preferences to guide customized slide generation. We also introduce a novel chain-of-speech mechanism to align slide content with planned oral narration. Such a design significantly enhances the quality of generated slides and enables downstream applications like video presentations. To support this new task, we construct a benchmark dataset that captures diverse user preferences, with meticulously designed interpretable metrics for robust evaluation. Experiments demonstrate the effectiveness of our proposed approach. Code and data will be released upon paper publication.

SlideTailor: Personalized Presentation Slide Generation for Scientific Papers

A major challenge in developing robust and generalizable Human Activity Recognition (HAR) systems for smart homes is the lack of large and diverse labeled datasets. Variations in home layouts, sensor configurations, and individual behaviors further exacerbate this issue. To address this, we leverage the idea of embodied AI agents—virtual agents that perceive and act within simulated environments guided by internal world models. We introduce AgentSense, a virtual data generation pipeline in which agents live out daily routines in simulated smart homes, with behavior guided by Large Language Models (LLMs). The LLM generates diverse synthetic personas and realistic routines grounded in the environment, which are then decomposed into fine-grained actions. These actions are executed in an extended version of the VirtualHome simulator, which we augment with virtual ambient sensors that record the agents’ activities. Our approach produces rich, privacy-preserving sensor data that reflects real-world diversity. We evaluate AgentSense on five real HAR datasets. Models pretrained on the generated data consistently outperform baselines, especially in low-resource settings. Furthermore, combining the generated virtual sensor data with a small amount of real data achieves performance comparable to training on full real-world datasets. These results highlight the potential of using LLM-guided embodied agents for scalable and cost-effective sensor data generation in HAR.

AgentSense: Virtual Sensor Data Generation Using LLM Agents in Simulated Home Environments

Multimodal Large Language Models (MLLMs) achieve impressive performance once optimized on massive datasets. Such datasets often contain sensitive or copyrighted content, raising significant data privacy concerns. Regulatory frameworks mandating the 'right to be forgotten' drive the need for machine unlearning. This technique allows for the removal of target data without resource-consuming retraining. However, while well-studied for text, visual concept unlearning in MLLMs remains underexplored. A primary challenge is precisely removing a target visual concept without disrupting model performance on related entities. To address this, we introduce AUVIC, a novel visual concept unlearning framework for MLLMs. AUVIC applies adversarial perturbations to enable precise forgetting. This approach effectively isolates the target concept while avoiding unintended effects on similar entities. To evaluate our method, we construct VCUBench. It is the first benchmark designed to assess visual concept unlearning in group contexts. Experimental results demonstrate that AUVIC achieves state-of-the-art target forgetting rates while incurs minimal performance degradation on non-target concepts.

AUVIC: Adversarial Unlearning of Visual Concepts for Multi-modal Large Language Models

Pedestrian trajectory prediction is critical for ensuring safety in autonomous driving, surveillance systems, and urban planning applications. While early approaches primarily focus on one-hop pairwise relationships, recent studies attempt to capture high-order interactions by stacking multiple Graph Neural Network (GNN) layers. However, these approaches face a fundamental trade-off: insufficient layers may lead to under-reaching problems that limit the model's receptive field, while excessive depth can result in prohibitive computational costs. We argue that an effective model should be capable of adaptively modeling both explicit one-hop interactions and implicit high-order dependencies, rather than relying solely on architectural depth. To this end, we propose ViTE (Virtual graph Trajectory Expert router), a novel framework for pedestrian trajectory prediction. ViTE consists of two key modules: a Virtual Graph that introduces dynamic virtual nodes to model long-range and high-order interactions without deep GNN stacks, and an Expert Router that adaptively selects interaction experts based on social context using a Mixture-of-Experts design. This combination enables flexible and scalable reasoning across varying interaction patterns. Experiments on three benchmarks (ETH/UCY, NBA, and SDD) demonstrate that our method consistently achieves state-of-the-art performance, validating both its effectiveness and practical efficiency.

ViTE: Virtual Graph Trajectory Expert Router for Pedestrian Trajectory Prediction

Many high-level multi-agent planning problems, such as multi-robot navigation and path planning, can be modeled with deterministic actions and observations. In this work, we focus on such domains and introduce the class of Deterministic Decentralized POMDPs (Det-Dec-POMDPs)—a subclass of Dec-POMDPs with deterministic transitions and observations given the state and joint actions. We then propose a practical solver, Iterative Deterministic POMDP Planning (IDPP), based on the classic Joint Equilibrium Search for Policies framework, specifically optimized to handle large-scale Det-Dec-POMDPs that existing Dec-POMDP solvers cannot handle efficiently.

Premium content

Next from AAAI 2026

Open-World Object Counting in Videos

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES