United States

We present a new formal framework for generalized planning (GP) based on the situation calculus extended with LTL constraints. The GP problem is specified by a first-order basic action theory whose models are the problem instances. This low-level theory is then abstracted into a high-level propositional nondeterministic basic action theory with a single model. A refinement mapping relates the two theories. LTL formulas are used to specify the temporally extended goals as well as assumed trace constraints. If all LTL trace constraints hold at the low level and the high-level model can simulate all the low-level models with respect to the mapping, we say that we have a temporally lifted abstraction. We prove that if we have such an abstraction and the agent has a strategy to achieve a LTL goal under some trace constraints at the abstract level, then there exists a refinement of the strategy to achieve the refinement of the goal at the concrete level. We use LTL synthesis to generate the strategy at the abstract level. We illustrate our approach by synthesizing a program that solves a data structure manipulation problem.

AAAI 2025

Situation Calculus Temporally Lifted Abstractions for Generalized Planning

action change and causality

We are pleased to announce the Thirty-Ninth AAAI Conference on Artificial Intelligence (AAAI-25), which will be held in Philadelphia, Pennsylvania at the Pennsylvania Convention Center from February 25 to March 4, 2025.

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-25 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.

### [Invited Speakers](https://aaai.org/conference/aaai/aaai-25/aaai-25-invited-speakers/)

Register [here](https://aaai.org/conference/aaai/aaai-25/registration/)

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-25 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.



technical paper

Signed Distance Functions (SDFs) are a vital kind of implicit representations to represent high fidelity 3D surfaces. Current methods mainly leverage a neural network to learn an SDF from various supervisions including signed distances, 3D point clouds, or multi-view images. However, due to various reasons including the bias of neural network on low frequency content, 3D unaware sampling, sparsity in point clouds, or low resolutions of images, neural implicit representations still struggle to represent geometries with high frequency components like sharp structures, especially for the ones learned from images or point clouds. To overcome this challenge, we introduce a method for sharpening an low frequency SDF observation for sharper and more complete surfaces by recovering its high frequency components. Our key idea is to learn a mapping from a low frequency observation to a full frequency coverage in a data-driven manner, leading to a prior knowledge of shape consolidation in the frequency domain, dubbed frequency consolidation priors. To better generalize a learned prior to unseen shapes, we introduce to represent frequency components as embeddings and disentangle the embedding for low frequency component from the embedding for the full frequency component. This disentanglement allows the prior to simply recover the full frequency embedding by a test-time self-reconstruction on an unseen low frequency observation. Our evaluations under widely used benchmarks or real scenes show that our method can recover high frequency component and produce more accurate surfaces than the latest methods.

Sharpening Neural Implicit Functions with Frequency Consolidation Priors

Heterogeneous Graph Neural Networks (HGNNs) have achieved state-of-the-art performance in classifying molecular graphs, capitalizing on their ability to capture rich semantics. However, HGNNs for molecule property prediction exhibit significant susceptibility to adversarial attacks—a challenge that prior research has entirely overlooked. To fill this gap, this paper introduces the first study focused on robust graph-level representation learning tailored for heterogeneous molecular graphs. To achieve this goal, we propose a comprehensive $\textbf{R}$obust $\textbf{H}$eterogeneous $\textbf{G}$raph $\textbf{C}$lassification ($\textbf{RHGC}$) framework grounded in the Information Bottleneck principle, which aims to identify the most informative and least noisy heterogeneous subgraphs to derive robust, holistic representations.  This is specifically accomplished through a dedicated Node Semantic Purifier, which enhances node-level and semantic-level robustness by eliminating label-irrelevant interference using graph stochastic attention and the Hilbert-Schmidt Independence Criterion, along with a Global Graph Disentanglement method, which improves graph-level robustness by addressing information leak. Experiments on three molecular benchmarks demonstrate that RHGC enhances accuracy by an average of 5.06% under all three attack settings and meanwhile by 4.33% on clean data. Our code and model will be made publicly available, which can be found in the supplementary material.

Robust Heterogeneous Graph Classification for Molecular Property Prediction with Information Bottleneck

We study the problem of designing replication-proof bandit mechanisms when agents strategically register or replicate their own arms to maximize their payoff.
Specifically, we consider Bayesian agents who only know the distribution from which their own arms' mean rewards are sampled, which is the first to study Bayesian extension of Shin et al. 2022.
With Bayesian agents, in stark contrast to the previous work, analyzing the replication-proofness of an algorithm becomes significantly complicated even in a single-agent setting.
We provide sufficient and necessary conditions for an algorithm to be replication-proof in the single-agent setting, and prove that exploration-then-commit algorithm satisfies these properties.
These results centers around several analytical theorems in comparing the expected regret of multiple bandit instances, which might be of independent interest and never have been studied to the best of our knowledge. 
We expand this result to the multi-agent setting, and provide a replication-proof algorithm for any problem instance.
We finalize our result by proving its sublinear regret upper bound that matches H-UCB.

Replication-proof Bandit Mechanism Design with Bayesian Agents

Proof systems can be used for certification of logic problems, and proof complexity can inform us how succinct certificates can be.  

In the PSPACE complete logic QBF (Quantified Boolean Formulas)  refutation proofs often contain information that reproduce the witnesses of the quantified variables. This is known as strategy extraction. There are two known kinds of strategy extraction for proof systems, local strategy extraction and round-based strategy extraction. Formalisation of local strategy extraction was done previously, in this paper we formalise round-based strategy extraction. 

By formalising the strategy extraction into circuits we can show new p-simulations. P-simulations are processes that allow you to transform proofs from a weaker proof system to a stronger proof system. Thus we solve an open problem in QBF proof complexity that Extended QBF Frege p-simulates LD-Q(\Drrs)-Resolution. LD-Q(\Drrs)-Resolution is the underlying proof system for the solver Qute.

This is a positive result for certification. By clarifying the hierarchy of proof systems further suggests the feasibility of using known formats such as Extended QU-Resolution or QRAT to certify QCDCL solvers.

The p-simulation is our main result, but we also make other observations from the specifics of the formalisation.

Proof Simulation via Round-based Strategy Extraction for QBF

Point cloud salient object detection has attracted the attention of researchers in recent years. Since existing works do not fully utilize the geometry context of 3D objects, blurry boundaries are generated when segmenting objects with complex backgrounds. In this paper, we propose a geometry-aware 3D salient object detection network that explicitly clusters points into superpoints to enhance the geometric boundaries of objects, thereby segmenting complete objects with clear boundaries. Specifically, we first propose a simple yet effective superpoint partition module to cluster points into superpoints. In order to improve the quality of superpoints, we present a point cloud class-agnostic loss to learn discriminative point features for clustering superpoints from the object. After obtaining superpoints, we then propose a geometry enhancement module that utilizes superpoint-point attention to aggregate geometric information into point features for predicting the salient map of the object with clear boundaries. Extensive experiments show that our method achieves new state-of-the-art performance on the PCSOD dataset.

Geometry-Aware 3D Salient Object Detection Network

3D Change Detection (3DCD) has gradually become another research hotspot after image change detection. Recent works focus on using artificial labels for supervised or weakly-supervised training of siamese networks to segment change points. However, labeling every points of multi temporal point clouds is very expensive and time-consuming. In addition, these works lack effective self-supervised signals, and existing self-supervised signals often fail to capture sufficiently rich change information. To solve this problem, we assume that the powerful representation of 3D objects should model the consistency information of unchanged regions and be distinguishable from other objects. Based on this assumption, we propose a new unsupervised framework that learns multi-temporal point clouds change information through bidirectional optimization of change segmentor and feature extractor. In the training stage, we first design a foreknowledge point contrastive loss based on the characteristics of the 3DCD task to initialize the feature extractor well, and then propose a masked consistency loss to further learn the shared geometric information of unchanged regions in the multi-temporal point clouds, utilizing it as a free and powerful supervised signal to train a change segmentor. In the inference stage, only the segmentor is used to take multi-temporal point clouds as input and produce change segmentation result. A large number of experiments have been conducted on SLPCCD and Urb3DCD, two real-world datasets of streets and urban buildings, to verify that our proposed unsupervised method is highly competitive and even outperforms supervised methods in scenes where semantic information changes occur, exhibiting better performance in generalization ability and robustness.

MUCD: Unsupervised Point Cloud Change Detection via Masked Consistency

To unbiasedly evaluate multiple target policies, the dominant approach among RL practitioners is to run and evaluate each target policy separately. However, this evaluation method is far from efficient because samples are not shared across policies, and running target policies to evaluate themselves is actually not optimal. In this paper, we address these two weaknesses by designing a tailored behavior policy to reduce the variance of estimators across all target policies. Theoretically, we prove that executing this behavior policy with manyfold fewer samples outperforms on-policy evaluation on every target policy under characterized conditions. Empirically, we show our estimator has a substantially lower variance compared with previous best methods and achieves state-of-the-art performance in a broad range of environments.

Efficient Multi-Policy Evaluation for Reinforcement Learning

Neural surface representation has demonstrated remarkable success in the areas of novel view synthesis and 3D reconstruction. However, assessing the geometric quality of 3D reconstructions in the absence of ground truth mesh remains a significant challenge, due to its rendering-based optimization process and entangled learning of appearance and geometry with photometric losses. 
In this paper, we present a novel framework, i.e, GURecon, which establishes a geometric uncertainty field for the neural surface based on geometric consistency.
Different from existing methods that rely on rendering-based measurement, 
GURecon models a continuous 3D uncertainty field for the reconstructed surface, and is learned by an online distillation approach without introducing real geometric information for supervision. 
Moreover, in order to mitigate the interference of illumination on geometric consistency, a decoupled field is learned and exploited to finetune the uncertainty field. 
Experiments on various datasets demonstrate the superiority of GURecon in modeling 3D geometric uncertainty, as well as its plug-and-play extension to various neural surface representations and improvement on downstream tasks such as incremental reconstruction.

GURecon: Learning Detailed 3D Geometric Uncertainties for Neural Surface Reconstruction

Depression can be reflected by long-term human spatio-temporal facial behaviours. While human face videos recorded in real-world usually have long and variable lengths, existing video-based depression assessment approaches frequently re-sample/down-sample such videos to short and equal-length videos, or split each video into several equal-length segments, where segment-level spatio-temporal facial behaviours are suppressed as a vector-style representations for RNN-based long-term (video-level) modelling. Both strategies lead to crucial information loss and distortion. In this paper, we propose a novel graph-style data structure called Matrixial Graph and an effective Matrixial Graph Neural Network (MGNN) for face video-based depression assessment, which can directly and end-to-end model long-term depression-specific spatio-temporal facial cues from variable-length videos without resampling/splitting videos or suppressing video segments to vectors. Importantly, the nodes in our matrixial graph are capable of including matrices of different shapes, and thus nodes of a matrix graph can directly represent all frame-level 2D facial feature maps (or images themselves) of an entire video regardless of its length. Then, our MGNN is the first GNN that can jointly process matrixial graphs containing varying numbers of nodes, which further learns matrix-style edge features, thereby facilitating to explicit model video-level multi-scale spatio-temporal facial behaviours among matrixial graph nodes for depression assessment. Experiments show that the explicit spatio-temporal modeling on 2D facial feature maps, facilitated by our matrixial graph/MGNN, provided significant benefits, leading our approach to achieve new state-of-the-art performances on AVEC2013 and AVEC2014 datasets with large advantages. Our code is provided in the Supplementary Material.

DepMGNN: Matrixial Graph Neural Network for Video-based Automatic Depression Assessment

Causal language models acquire vast amount of knowledge from general text corpus during pretraining, but the efficiency of knowledge learning is known to be unsatisfactory, especially when learning from knowledge-dense and small-sized corpora. The deficiency can come from long-distance dependencies which are hard to capture by language models, and overfitting to co-occurrence patterns and distracting clues in the training text. To address these issues, the paper proposes a method to enhance knowledge learning during language model pretraining, by enhancing elusive but important clues in text discovered by the language model themselves. We found that larger language models pay more attention to non-obvious but important clues, which are often overlooked by smaller language models. Therefore, we can identify these clues by contrasting the attention weights of large and small language models. We use the identified clues as a guide to perform token-dropout data augmentation on the training text, and observed a significant boost in both small and large models' performance in fact memorization. This shows that the behavior contrast between more and less-performant language models contains important clues for knowledge learning, and it can be "amplified" for a straight-forward improvement in knowledge learning efficiency.

Downloads

Next from AAAI 2025

Sharpening Neural Implicit Functions with Frequency Consolidation Priors

Similar lecture

A Syntactic Approach to Computing Complete and Sound Abstraction in the Situation Calculus