Singapore

Spatio-temporal alignment is crucial for temporal modeling of end-to-end (E2E) perception in autonomous driving (AD), providing valuable structural and textural prior information.
Existing methods typically rely on the attention mechanism to align objects across frames, simplifying the motion model with a unified explicit physical model (constant velocity, etc.). 
These approaches prefer semantic features for implicit alignment, challenging the importance of explicit motion modeling in the traditional perception paradigm. 
However, variations in motion states and object features across categories and frames render this alignment suboptimal.
To address this, we propose HAT, a spatio-temporal alignment module that allows each object to adaptively decode the optimal alignment proposal from multiple hypotheses without direct supervision.
Specifically, HAT first utilizes multiple explicit motion models to generate spatial anchors and motion-aware feature proposals for historical instances. 
It then performs multi-hypothesis decoding by incorporating semantic and motion cues embedded in cached object queries, ultimately providing the optimal alignment proposal for the target frame.
On nuScenes, HAT consistently improves 3D temporal detection and tracking performance across diverse baselines.
It achieves state-of-the-art tracking results with 46.0\% AMOTA on the test set when paired with DETR3D detector.
In an object-centric E2E AD method, HAT enhances perception accuracy (+1.3\% mAP, +3.1\% AMOTA) and reduces the collision rate by 32\%.
When semantics are corrupted (nuScenes-C), the enhancement of motion modeling by HAT enables more robust perception in the E2E AD framework.

AAAI 2026

Rethinking the Spatio-Temporal Alignment of End-to-End 3D Perception

cv: motion & tracking

cv: vision for robotics & autonomous driving

cv: object detection & categorization

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.

The well-known Condorcet Jury Theorem states that, under majority rule, the better of two alternatives is chosen with probability approaching one as the population grows. We study an asymmetric setting where voters face varying participation costs and share a (possibly heuristic) belief about their ability to influence the outcome, aka, pivotality.

In a costly voting setup where voters abstain if their participation cost is more than their pivotality estimate, we identify a single property of the heuristic belief---weakly vanishing pivotality---that gives rise to multiple stable equilibria in which elections are nearly tied. In contrast, strongly vanishing pivotality (as in the standard Calculus of Voting model) yields a unique, trivial equilibrium where only zero-cost voters participate as the population grows. We then characterise when nontrivial equilibria satisfy a version of the Jury Theorem: below a sharp threshold, the majority-preferred candidate wins with probability approaching one; above it, both candidates either win with equal probability or maintain a constant winning chance, independent of population size or participation cost distribution.

On Condorcet’s Jury Theorem with Abstention

Training and deploying multiple vision transformer (ViT) models for different resource constraints is costly and inefficient. To address this, we propose transforming a pre-trained ViT into a stratified knowledge-density super-network, where knowledge is hierarchically organized across weights. This enables flexible extraction of sub-networks that retain maximal knowledge for varying model sizes. We introduce Weighted PCA for Attention Contraction (WPAC), which concentrates knowledge into a compact set of critical weights. WPAC applies token-wise weighted principal component analysis to intermediate features and injects the resulting transformation and inverse matrices into adjacent layers, preserving the original network function while enhancing knowledge compactness. To further promote stratified knowledge organization, we propose Progressive Importance-Aware Dropout (PIAD). PIAD progressively evaluates the importance of weight groups, updates an importance-aware dropout list, and trains the super-network under this dropout regime to promote knowledge stratification. Experiments demonstrate that WPAC outperforms existing pruning criteria in knowledge concentration, and the combination with PIAD offers a strong alternative to state-of-the-art model compression and model expansion methods.

Stratified Knowledge-Density Super-Network for Scalable Vision Transformers

Knowledge distillation (KD) is a widely adopted technique for transferring the capabilities of large teacher models to smaller student models, thereby significantly reducing inference costs and memory consumption. However, existing KD methods are all constrained by an inherent greedy optimization objective, rooted in the assumption of teacher superiority: "Trust all teacher-generated outputs (TGOs)" and "Distrust any student-generated outputs (SGOs) unsupported by the teacher". We propose ASKD, a novel KD method with adaptive skewness determined by sample quality, refining this objective to: "Learn TGOs proportionally to their quality, and distrust only low-quality unsupported SGOs". ASKD comprises three key components: (1) A reinforcement learning-style optimization formulation to mitigate the inherent approximation bias in sample-based Kullback-Leibler (KL) divergence approximations used by previous KD methods; (2) Well-designed quality supervision signals to map and achieve adaptive skewness in skewed KL loss, pioneering the usage of sample quality to adjust learning magnitudes; (3) A gradient-clip function on high-quality SGOs for findings that high-quality SGOs in KL loss fail to yield positive updates and even cause adverse effects on some samples. Extensive experiments indicate that ASKD builds high-performance student models across various tasks, including instruction following, mathematical reasoning, and code generation, outperforming state-of-the-art methods comprehensively and surpassing GRPO-like approaches that use advantages as multiplicative factors. We also provide detailed mathematical proofs demonstrating properties such as Lipschitz continuity of the update coefficient and uniform convergence of the loss function, ensuring theoretical rigor for key components of ASKD.

ASKD: Reinforcement Learning-Style Knowledge Distillation with Quality-Adaptive Skewness

Defending large language models (LLMs) against jailbreak attacks is crucial for ensuring their safe deployment. Existing defense strategies typically rely on predefined static criteria to differentiate between harmful and benign prompts. However, such rigid rules fail to accommodate the inherent complexity and dynamic nature of real-world jailbreak attacks. In this paper, we focus on the novel challenge of adaptive defense against diverse jailbreaks. We propose a new concept "mirror'', which is a dynamically generated prompt that reflects the syntactic structure of the input while ensuring semantic safety. The discrepancies between input prompts and their corresponding mirrors serve as guiding principles for defense. A novel defense model, MirrorShield, is further proposed to detect and calibrate risky inputs based on the crafted mirrors. Evaluated on multiple benchmark datasets and compared against ten state-of-the-art attack methods, MirrorShield demonstrates superior defense performance and promising generalization capabilities.

MirrorShield: Towards Dynamic Adaptive Defense Against Jailbreaks via Entropy-Guided Mirror Crafting

Reconstructing dynamic scenes has long been a challenging task in 3D vision. Previous mainstream methods based on 3D Gaussian Splatting typically employ a single deformation field to directly model spatiotemporal changes. However, such one-step deformation struggles to capture diverse and complex motion patterns. To address this limitation, we propose decomposing the one-step deformation into a multi-step process, where each step is represented by a deformation layer. Additionally, we introduce a weight prediction mechanism for each layer to control the extent of deformation at every step. We provide two types of deformation layers based on implicit and explicit approaches. Moreover, while the deformation layer is time-conditioned, the Gaussians' behavior may still be influenced by their time-invariant properties. Therefore, we propose a fully time-agnostic scale modulation block to modulate the scaling changes of Gaussians. Extensive experiments on D-NeRF, Neu3D, and HyperNeRF demonstrate that our method achieves state-of-the-art performance.

Multi-Step Deformable Gaussian Splatting for Dynamic Scene Rendering

Meme is an expressive medium that often conveys rich emotions and intentions. Recent studies have confirmed the critical role of metaphors in meme understanding. However, existing metaphor research heavily relies on manual annotations, and mainstream vision-language models (VLMs) still struggle with the recognition and comprehension of metaphors. To address these challenges, we introduce MetaGPT, the first vision-language model specifically designed for meme metaphor understanding. MetaGPT is capable of identifying and extracting metaphors in memes, and generating accurate meme interpretations. Furthermore, we construct a dedicated dataset for meme understanding, MUnd, which comprises approximately 32,000 high-quality question-answer (QA) pairs across three core tasks: metaphor detection, metaphor domain extraction, and meme interpretation. Based on MUnd, we further propose an evaluation benchmark for meme understanding and conduct a comprehensive assessment of existing VLMs. Experimental results reveal that current models still face challenges in metaphor comprehension, while MetaGPT consistently outperforms them across all tasks, highlighting its potential in advancing meme understanding.

MetaGPT: A Large Vision-Language Model for Meme Metaphor Understanding

Vision-Language Models (VLMs) have demonstrated remarkable capabilities in visual classification tasks. Existing methods for enhancing VLMs on this task often rely heavily on direct category-to-image matching, which limits generalization and results in suboptimal performance. In addition, these methods provide no understanding of why a specific category is chosen. To address these limitations, we introduce a new deliberative visual classification task that decomposes the classification process into multiple deliberative steps and leverages Large Language Models (LLMs) to perform explicit reasoning before the final decision. Specifically, we propose a Retrieval-driven Reasoning model (RdR) with two components, i.e., retrieval database construction and deliberative category prediction. The first component leverages LLMs to extract category-relevant descriptors and constructs a retrieval database for effective image–descriptor matching. The second component facilitates multiple deliberative steps and performs explicit reasoning based on the retrieved descriptors to augment the category prediction. Extensive experiments on multiple datasets demonstrate that RdR consistently outperforms strong baselines, highlighting its robustness and generalization ability.

Retrieval-driven Reasoning for Deliberative Visual Classification

Information retrieval (IR) systems play a critical role in navigating information overload across various applications. Existing IR benchmarks primarily focus on simple queries that are semantically analogous to single- and multi-hop relations, overlooking complex logical queries involving first-order logic operations such as conjunction (∧), disjunction (∨), and negation (¬). 
Thus, these benchmarks can not be used to sufficiently evaluate the performance of IR models on complex queries in real-world scenarios. To address this problem, we propose a novel method leveraging large language models (LLMs) to construct a new IR dataset ComLQ for Complex Logical Queries, which comprises 2,909 queries and 11,251 candidate passages. A key challenge in constructing the dataset lies in capturing the underlying logical structures within unstructured text. Therefore, by designing the subgraph-guided prompt with the subgraph indicator, an LLM (such as GPT-4o) is guided to generate queries with specific logical structures based on selected passages. All query-passage pairs in ComLQ are ensured structure conformity and evidence distribution through expert annotation. To better evaluate whether retrievers can handle queries with negation, we further propose a new evaluation metric, Log-Scaled Negation Consistency (LSNC@K). As a supplement to standard relevance-based metrics (such as nDCG and mAP), LSNC@K measures whether top-K retrieved passages violate negation conditions in queries. Our experimental results under zero-shot settings demonstrate existing retrieval models' limited performance on complex logical queries, especially on queries with negation, exposing their inferior capabilities of modeling exclusion. In summary, our ComLQ offers a comprehensive and fine-grained exploration, paving the way for future research on complex logical queries in IR.

ComLQ: Benchmarking Complex Logical Queries in Information Retrieval

Large Language Models (LLMs) have demonstrated significant potential across various domains. However, they often struggle with integrating external knowledge and performing complex reasoning, leading to hallucinations and unreliable outputs. Retrieval Augmented Generation (RAG) has emerged as a promising paradigm to mitigate these issues by incorporating external knowledge. Yet, conventional RAG approaches—especially those based on vector similarity—fail to effectively handle relational structures and multi-step reasoning. In this work, we propose CogGRAG, a human cognition inspired, graph-based RAG framework designed for Knowledge Graph Question Answering (KGQA). CogGRAG mimics human reasoning through a three-stage process: (1) top-down problem decomposition via mind map construction; (2) structured retrieval of local and global knowledge from external Knowledge Graphs (KGs); and (3) bottom-up reasoning with self-verification. Unlike previous tree-based decomposition methods such as MindMap or Graph-CoT, CogGRAG unifies the entire reasoning process under a global mind map with early-stage, graph-structured retrieval and integrates dual-process verification to mitigate error propagation. Extensive experiments demonstrate that CogGRAG outperforms existing methods in both accuracy and reliability. We provide our code and data here: https://anonymous.4open.science/r/RAG-5883.

Human Cognition Inspired RAG with Knowledge Graph for Complex Problem Solving

While scaling the length of responses at test-time has been shown to markedly improve the reasoning abilities and performance of large language models (LLMs), it often results in verbose outputs and increases inference cost. Prior approaches for efficient test-time scaling, typically using universal budget constraints or query-level length optimization, do not leverage historical information from previous encounters with the same problem during training. We hypothesize that this limits their ability to progressively make solutions more concise over time. To address this, we present History-Aware Policy Optimization (HAPO), which keeps track of a history state (e.g., the minimum length over previously generated correct responses) for each problem. HAPO employs a novel length reward function based on this history state to incentivize the discovery of correct solutions that are more concise than those previously found. Crucially, this reward structure avoids overly penalizing shorter incorrect responses with the goal of facilitating exploration towards more efficient solutions. By combining this length reward with a correctness reward, HAPO jointly optimizes for correctness and efficiency. We use HAPO to train DeepSeek-R1-Distill-Qwen-1.5B, DeepScaleR-1.5B-Preview, and Qwen-2.5-1.5B-Instruct, and evaluate HAPO on several math benchmarks that span various difficulty levels. Experiment results demonstrate that HAPO effectively induces LLMs’ concise reasoning abilities, producing length reductions of 33-59% with accuracy drops of only 2-5%.

Content not yet available

Next from AAAI 2026

On Condorcet’s Jury Theorem with Abstention

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES