Singapore

Previous methods evaluate reward models by testing them on a fixed pairwise ranking test set, but they typically do not provide performance information on each preference dimension. In this work, we address the evaluation challenge of reward models by probing preference representations. To confirm the effectiveness of this evaluation method, we construct a Multi-dimensional Reward Model Benchmark (MRMBench), a collection of six probing tasks for different preference dimensions. We design it to favor and encourage reward models that better capture preferences across different dimensions. Furthermore, we introduce an analysis method, inference-time probing, which identifies the dimensions used during the reward prediction and enhances its interpretability. Through extensive experiments, we find that MRMBench strongly correlates with LLM alignment performance, supporting it as a reliable reference for developing advanced reward models. By analyzing the evaluation results on MRMBench, we reveal that reward models struggle to simultaneously capture preferences across multiple dimensions, highlighting the potential of multi-objective optimization in reward modeling. Furthermore, our results demonstrate that the proposed inference-time probing method provides a reliable metric for assessing the confidence of reward predictions, leading to improved alignment of large language models.

AAAI 2026

Probing Preference Representations: A Multi-Dimensional Evaluation and Analysis Method for Reward Models

inference-time probing

probing reward model

reward model interpretability

reward model evaluation

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.

poster

Multi-modal knowledge graph completion (MMKGC) aims to infer missing entities of triples by leveraging heterogeneous information in knowledge graph (KG). However, existing approaches often struggle with inconsistent modality alignment, limited reasoning depth, and insufficient negative sample quality. In this work, we propose HFR-MKGC, a novel framework that integrates hierarchical modal fusion and Multimodal Large Language Model (MLLM) reasoning for robust and expressive MMKGC. Specifically, we introduce a relation-guided hierarchical modal fusion module, which conducts fine-grained intra-visual fusion and relation-guided cross-modal integration to yield rich entity representations. HFR-MKGC employs a fine-tuned MLLM to perform instruction-based triple reasoning, producing candidate entities for completion. Then, it constructs hard negative samples through textual perturbation by MLLM and visual feature augmentation with rotation and noise.HFR-MKGC optimizes the model via adversarial training. Extensive experiments on three MMKGC benchmarks demonstrate that our method outperforms state-of-the-art methods, validating its effectiveness in MMKGC.

HFR-MKGC: Hierarchical Fusion Reasoning with MLLMs for Multi-modal Knowledge Graph Completion

In Split Federated Learning (SFL), the clients collaboratively train a model with the help of a server by splitting the model into two parts. Part-1 is trained locally at each client and aggregated by the aggregator at the end of each round. Part-2 is trained at a server that sequentially processes the intermediate activations received from each client. We study the phenomenon of catastrophic forgetting (CF) in SFL in the presence of data heterogeneity. In detail, due to the nature of SFL, local updates of part-1 may drift away from global optima, while part-2 is sensitive to the processing sequence, similar to forgetting in continual learning (CL). Specifically, we observe that the trained model performs better in classes (labels) seen at the end of the sequence. We investigate this phenomenon with emphasis on key aspects of SFL, such as the processing order at the server and the cut layer. Based on our findings, we propose Hydra, a novel mitigation method inspired by multi-head neural networks and adapted for the SFL's setting. Extensive numerical evaluations show that Hydra outperforms baselines and methods from the literature.

Data Heterogeneity and Forgotten Labels in Split Federated Learning

Emotional support is a core capability in human-AI interaction, with applications including psychological counseling, role play, and companionship. However, existing evaluations of large language models (LLMs) often rely on short, static dialogues and fail to capture the dynamic and long-term nature of emotional support. To overcome this limitation, we shift from snapshot-based evaluation to trajectory-based assessment, adopting a user-centered perspective that evaluates models based on their ability to improve and stabilize user emotional states over time. Our framework constructs a large-scale benchmark consisting of 328 emotional contexts and 1,152 disturbance events, simulating realistic emotional shifts under evolving dialogue scenarios. To encourage psychologically grounded responses, we constrain model outputs using validated emotion regulation strategies such as situation selection and cognitive reappraisal. User emotional trajectories are modeled as a first-order Markov process, and we apply causally-adjusted emotion estimation to obtain unbiased emotional state tracking. Based on this framework, we introduce three trajectory-level metrics: Baseline Emotional Level (BEL), Emotional Trajectory Volatility (ETV), and Emotional Centroid Position (ECP). These metrics collectively capture user emotional dynamics over time and support comprehensive evaluation of long-term emotional support performance of LLMs. Extensive evaluations across a diverse set of LLMs reveal significant disparities in emotional support capabilities and provide actionable insights for model development.

Detecting Emotional Dynamic Trajectories: An Evaluation Framework for Emotional Support in Language Models

Fracture injuries often lead to complex bone fragmentations, posing significant challenges for accurate segmentation in surgical planning and trauma assessment. Manual annotation of each fragment is time-consuming and inconsistent, while existing automated methods often fail to separate individual fragments due to the wide variation in fracture types, irregular fracture surface, and close inter-fragment contact. 
To address these challenges, we introduce FracSegmentator, a deep learning approach for bone fragment instance segmentation. The model takes extracted bone regions in CT as input and isolates individual fragments by identifying fracture surfaces and separating closely contacting structures. Central to our approach is a Trauma-Prior-Guided Contrastive Learning module, which incorporates clinical knowledge through memory-based attention to better distinguish fractured surfaces from healthy regions. 
We evaluate FracSegmentator on four datasets that cover a range of anatomical sites and fracture patterns. The method achieves state-of-the-art results across all datasets and demonstrates strong generalization capabilities. By delivering accurate and efficient fragment-level segmentation, FracSegmentator supports critical downstream tasks such as automated fracture diagnosis, surgical planning, and preoperative reduction simulation.

FracSegmentator: Fracture Instance Segmentation with Trauma-Prior-Guided Contrastive Learning

Due to the diversity of brain anatomy and the scarcity of annotated data, supervised anomaly detection for brain MRI remains challenging, driving the development of unsupervised anomaly detection (UAD) approaches. 
Current UAD methods typically utilize synthetically generated noise perturbations on healthy MRIs to train generative models for normal anatomy reconstruction, enabling anomaly detection via residual mapping.
However, such simulated anomalies lack the biophysical fidelity and morphological complexity characteristic of true clinical lesions.
To advance UAD in brain MRI, we conduct the first systematic frequency-domain analysis of pathological signatures, revealing two key properties: (1) anomalies exhibit unique frequency patterns distinguishable from normal anatomy, and (2) low-frequency signals maintain consistent representations across healthy scans. These insights motivate our Frequency-Decomposition Preprocessing (FDP) framework—the first UAD method to leverage frequency-domain reconstruction for simultaneous pathology suppression and anatomical preservation. FDP can integrate seamlessly with existing anomaly simulation techniques, consistently enhancing detection performance across diverse architectures while maintaining diagnostic fidelity.
Experimental results demonstrate that FDP consistently improves anomaly detection performance when integrated with existing methods. Notably, FDP achieves a $17.63\%$ increase in DICE score with LDM while maintaining robust improvements across multiple baselines.

FDP: A Frequency-Decomposition Preprocessing Pipeline for Unsupervised Anomaly Detection in Brain MRI

Vision-Language-Action (VLA) models process visual inputs independently at each timestep, discarding valuable temporal information inherent in robotic manipulation tasks. This frame-by-frame processing makes models vulnerable to visual noise while ignoring the substantial coherence between consecutive frames in manipulation sequences. We propose Temporal Token Fusion (TTF), a training-free approach that intelligently integrates historical and current visual representations to enhance VLA inference quality. Our method employs dual-dimension detection combining efficient grayscale pixel difference analysis with attention-based semantic relevance assessment, enabling selective temporal token fusion through hard fusion strategies and keyframe anchoring to prevent error accumulation. Comprehensive experiments across LIBERO, SimplerEnv, and real robot tasks demonstrate consistent improvements: 4.0 percentage points average on LIBERO (72.4\% vs 68.4\% baseline), cross-environment validation on SimplerEnv (4.8\% relative improvement), and 8.7\% relative improvement on real robot tasks. Our approach proves model-agnostic, working across OpenVLA and VLA-Cache architectures. Notably, TTF reveals that selective Query matrix reuse in attention mechanisms enhances rather than compromises performance, suggesting promising directions for direct KQV matrix reuse strategies that achieve computational acceleration while improving task success rates.

TTF-VLA: Temporal Token Fusion via Pixel-Attention Integration for Vision-Language-Action Models

Differentially Private Stochastic Gradient Descent (DPSGD) is widely used to protect sensitive data during the training of machine learning models, but its privacy guarantee often comes at a large cost of model performance due to the lack of tight theoretical bounds quantifying privacy loss. While recent efforts have achieved more accurate privacy guarantees, they still impose some assumptions prohibited from practical applications, such as convexity and complex parameter requirements, and rarely investigate in-depth the impact of privacy mechanisms on the model’s utility. In this paper, we provide a rigorous privacy characterization for DPSGD with general $L$-smooth and non-convex loss functions, revealing converged privacy loss with iteration in bounded-domain cases. Specifically, we track the privacy loss over multiple iterations, leveraging the noisy smooth-reduction property, and further establish comprehensive convergence analysis in different scenarios. In particular, we show that for DPSGD with a bounded domain, (i) the privacy loss can still converge without the convexity assumption, (ii) a smaller bounded diameter can improve both privacy and utility simultaneously under certain conditions, and (iii) the attainable big-$\mathcal{O}$ order of the privacy utility trade-off for DPSGD with gradient clipping (DPSGD-GC) and for DPSGD-GC with bounded domain (DPSGD-DC) and $\mu$-strongly convex population risk function, respectively. Experiments via membership inference attack (MIA) in a practical setting validate insights gained from the theoretical results.

An Improved Privacy and Utility Analysis of Differentially Private SGD with Bounded Domain and Smooth Losses

Recent advancements in Large Reasoning Models (LRMs), such as OpenAI's o1/o3 and DeepSeek-R1, have demonstrated remarkable performance in specialized reasoning tasks through human-like deliberative thinking and long chain-of-thought reasoning. However, our systematic evaluation across various model families (DeepSeek, Qwen, and LLaMA) and scales (7B to 32B) reveals that acquiring these deliberative reasoning capabilities significantly reduces the foundational capabilities of LRMs, including notable declines in helpfulness and harmlessness, alongside substantially increased inference costs. Importantly, we demonstrate that adaptive reasoning---employing modes like Zero-Thinking, Less-Thinking, and Summary-Thinking---can effectively alleviate these drawbacks. Our empirical insights underline the critical need for developing more versatile LRMs capable of dynamically allocating inference-time compute according to specific task characteristics.

Trade-offs in Large Reasoning Models: An Empirical Analysis of Deliberative and Adaptive Reasoning over Foundational Capabilities

*Universal Safety Controllers (USCs)* are a promising logical control framework that guarantees the satisfaction of a given temporal safety specification when applied to any realizable plant model. Unlike traditional methods, which synthesize one logical controller over a given detailed plant model, USC synthesis constructs a *generic controller* whose outputs are conditioned by plant behavior, called *prophecies*. Thereby, USCs offer strong generalization and scalability benefits over classical logical controllers. However, the exact computation and verification of prophecies remain computationally challenging. 
In this paper, we introduce an approximation algorithm for USC synthesis that addresses these limitations via learning. Instead of computing exact prophecies, which reason about sets of trees via automata, we only compute under- and over-approximations from (small) example plants and infer computation tree logic (CTL) formulas as representations of prophecies. The resulting USC generalizes to unseen plants via a verification step and offers improved efficiency and explainability through small and concise CTL prophecies, which remain human-readable and interpretable. Experimental results demonstrate that our learned prophecies remain generalizable, yet are significantly more compact and interpretable than their exact tree automata representations.

Universal Safety Controllers with Learned Prophecies

The Shapley value is the prevalent solution for fair division problems in which a payout is to be divided among multiple agents. By adopting a game-theoretic view, the idea of fair division and the Shapley value can also be used in machine learning to quantify the individual contribution of features or data points to the performance of a predictive model.Despite its popularity and axiomatic justification, the Shapley value suffers from a computational complexity that scales exponentially with the number of entities involved, and hence requires approximation methods for its reliable estimation. We propose SVA$k_{\text{ADD}}$, a novel approximation method that fits a $k$-additive surrogate game. By taking advantage of $k$-additivity, we are able to elicit the exact Shapley values of the surrogate game and then use these values as estimates for the original fair division problem. The efficacy of our method is evaluated empirically and compared to competing methods.

Downloads

Next from AAAI 2026

HFR-MKGC: Hierarchical Fusion Reasoning with MLLMs for Multi-modal Knowledge Graph Completion

Similar lecture

GRAM-R²: Self-Training Generative Foundation Reward Models for Reward Reasoning

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from AAAI 2026

HFR-MKGC: Hierarchical Fusion Reasoning with MLLMs for Multi-modal Knowledge Graph Completion

Similar lecture

GRAM-R²: Self-Training Generative Foundation Reward Models for Reward Reasoning

Downloads