AAAI 2025 论文列表

Medical image segmentation plays an important role in clinical decision making, treatment planning, and disease tracking. However, it still faces two major challenges. On the one hand, there is often a "soft boundary" between foreground and background in medical images, with poor illumination and low contrast further reducing the distinguishability of foreground and background within the image. On the other hand, co-occurrence phenomena are widespread in medical images, and learning these features is misleading to the model's judgment. To address these challenges, we propose a general framework called Contrast-Driven Medical Image Segmentation (ConDSeg). First, we develop a contrastive training strategy called Consistency Reinforcement. It is designed to improve the encoder's robustness in various illumination and contrast scenarios, enabling the model to extract high-quality features even in adverse environments. Second, we introduce a Semantic Information Decoupling module, which is able to decouple features from the encoder into foreground, background, and uncertainty regions, gradually acquiring the ability to reduce uncertainty during training. The Contrast-Driven Feature Aggregation module then contrasts the foreground and background features to guide multi-level feature fusion and key feature enhancement, further distinguishing the entities to be segmented. We also propose a Size-Aware Decoder to solve the scale singularity of the decoder. It accurately locate entities of different sizes in the image, thus avoiding erroneous learning of co-occurrence features. Extensive experiments on five datasets across three scenarios demonstrate the state-of-the-art performance of our method, proving its advanced nature and general applicability to various medical image segmentation scenarios.

We present an algorithm for computing pure-strategy epsilon-perfect Bayesian equilibria in sequential auctions with continuous action and value spaces. Importantly, our algorithm includes a verification phase that computes an upper bound on the utility loss of the found strategies. Prior work on equilibrium computation in auctions with verification has focussed on the single-round case, but the methods do not work for sequential auctions because of two main challenges: (1) there are infinitely many subgames, and (2) the setting has no optimal substructure as bidders' beliefs and best response strategies depend on the strategies of previous rounds. We make two contributions. First, we introduce a tailor-made game abstraction that discretizes the auction and augments the state space with the public beliefs, such that an approximate equilibrium can be computed via dynamic programming. Second, we prove a decomposition theorem to upper bound the utility loss of the computed equilibrium. This is essential because it is neither guaranteed that the auction has an equilibrium nor that any algorithm converges to it. We validate our algorithm on multiple settings with known equilibria and apply it to a new multi-round combinatorial auction.

Strategic interactions can be represented more concisely, and analyzed and solved more efficiently, if we are aware of the symmetries within the multiagent system. Symmetries also have conceptual implications, for example for equilibrium selection. We study the computational complexity of identifying and using symmetries. Using the classical framework of normal-form games, we consider game symmetries that can be across some or all players and/or actions. We find a strong connection between game symmetries and graph automorphisms, yielding graph automorphism and graph isomorphism completeness results for characterizing the symmetries present in a game. On the other hand, we also show that the problem becomes polynomial-time solvable when we restrict the consideration of actions in one of two ways. Next, we investigate when exactly game symmetries can be successfully leveraged for Nash equilibrium computation. We show that finding a Nash equilibrium that respects a given set of symmetries is PPAD- and CLS-complete in general-sum and team games respectively---that is, exactly as hard as Brouwer fixed point and gradient descent problems. Finally, we present polynomial-time methods for the special cases where we are aware of a vast number of symmetries, or where the game is two-player zero-sum and we do not even know the symmetries.

There has been tremendous progress in the past decade in the field of quantified Boolean formulas (QBF), both in practical solving as well as in creating a theory of corresponding proof systems and their proof complexity analysis. Both for solving and for proof complexity, it is important to have interesting formula families on which we can test solvers and gauge the strength of the proof systems. There are currently few such formula families in the literature. We initiate a general programme how to transform computationally hard problems (located in the polynomial hierarchy) into QBFs hard for the main QBF resolution systems Q-Res and QU-Res that relate to core QBF solvers. We illustrate this general approach on three problems from graph theory and logic. This yields QBF families that are provably hard for Q-Res and QU-Res (without any complexity assumptions).

Developing competency in artificial intelligence is becoming increasingly crucial for computer science (CS) students at all levels of the CS curriculum. However, most previous research focuses on advanced CS courses, as traditional introductory courses provide limited opportunities to develop AI skills and knowledge. This paper introduces an introductory CS course where students learn computational thinking through computer vision, a sub-field of AI, as an application context. The course aims to achieve computational thinking outcomes alongside critical thinking outcomes that expose students to AI approaches and their societal implications. Through experiential activities such as individual projects and reading discussions, our course seeks to balance technical learning and critical thinking goals. Our evaluation, based on pre-and post-course surveys, shows an improved sense of belonging, self-efficacy, and AI ethics awareness among students. The results suggest that an AI-focused context can enhance participation and employability, student-selected projects support self-efficacy, and ethically grounded AI instruction can be effective for interdisciplinary audiences. Students' discussions on reading assignments demonstrated deep engagement with the complex challenges in today's AI landscape. Finally, we share insights on scaling such courses for larger cohorts and improving the learning experience for introductory CS students.

This talk explores the challenge of customizing large-scale AI models, particularly generative AI, on cost-effective devices with limited memory and energy resources. Modern AI models demand substantial computational power, often relying on specialized hardware such as GPUs. To address this, the talk introduces compression-aware computing, a framework enabling AI models to recognize and adapt to their compressed states while preserving performance. Compression-aware computing integrates compression techniques like sparsification, quantization, and low-rank decomposition to enhance the efficiency and accuracy of AI models, broadening these models' accessibility across diverse devices. Additionally, this talk highlights one rationale of scalable and sustainable AI in advancing Alzheimer’s research by facilitating the analysis of large single-cell transcriptomics datasets for gene-gene interaction discovery.

The success of 3D Gaussian Splatting (3DGS) in static scenes has inspired numerous attempts to construct Free-Viewpoint Videos (FVVs) of dynamic scenes from multi-view videos. Despite advancements in current techniques, simultaneously achieving photo-realistic view synthesis results, fast on-the-fly training, real-time rendering, and low storage costs remains a formidable problem. To address these challenges, we propose the first Gaussian-based streamable FVV intelligent compression framework named iFVC. Specifically, we utilize an anchor-based Gaussian representation to model the scene. To achieve on-the-fly training, we propose a Binary Transformation Cache (BTC) to model the dynamic changes between adjacent timesteps, which not only ensures compactness but also supports precise bit rate estimation. Furthermore, we carefully design a high-resolution transformation tri-plane assisted by a saliency grid as our BTC, allowing for accurate dynamic capture. The entire pipeline is regarded as a joint optimization of rate and distortion to achieve optimal compression performance. Experiments on widely used datasets demonstrate the state-of-the-art performance of our framework in both synthesis quality and efficiency, i.e., achieving per-frame training in 13 seconds with a storage cost of 0.1 MB and real-time rendering at 120 FPS.

Class-Incremental Learning (CIL) requires an artificial intelligence system to learn different tasks without class overlaps continually. To achieve CIL, some methods introduce the Pre-Trained Model (PTM) and leverage the generalized feature representation of PTM to learn downstream incremental tasks continually. However, the generalized feature representations of PTM are not adaptive and discriminative for these various incremental classes, which may be out of distribution for the pre-trained dataset. In addition, since the incremental classes cannot be learned at once, the class relationship cannot be constructed optimally, leading to undiscriminating feature representation for understream tasks. Thus, we propose a novel Pre-Trained Model-based Class-Incremental Learning (PTM-CIL) method to explore the potential of PTM and obtain optimal class relationships. Inspired by Neural Collapse theory, we introduce the frozen Equiangular Tight Frame classifier to construct optimal classifier structure for all seen classes, guiding the feature representation adaptation for downstream continual tasks. Specifically, Task-Related Adaptation is proposed to modulate the generalized feature representation to bridge the gap between the pre-trained dataset and various downstream datasets. Then, the Feature Compression Module is introduced to compress various features to the specific classifier weights, constructing the feature transfer pattern and satisfying the characteristic of Neural Collapse. Optimal Structural Alignment is designed to supervise the feature compression process, assisting in achieving optimal class relationships across different tasks. Sufficient experiments on seven datasets prove the effectiveness of our method.

Enabling models to recognize vast open-world categories has been a longstanding pursuit in object detection. By leveraging the generalization capabilities of vision-language models, current open-world detectors can recognize a broader range of vocabularies, despite being trained on limited categories. However, when the scale of the category vocabularies during training expands to a real-world level, previous classifiers aligned with coarse class names significantly reduce the recognition performance of these detectors. In this paper, we introduce Prova, a multi-modal prototype classifier for vast-vocabulary object detection. Prova extracts comprehensive multi-modal prototypes as initialization of alignment classifiers to tackle the vast-vocabulary object recognition failure problem. On V3Det, this simple method greatly enhances the performance among one-stage, two-stage, and DETR-based detectors with only additional projection layers in both supervised and open-vocabulary settings. In particular, Prova improves Faster R-CNN, FCOS, and DINO by 3.3, 6.2, and 2.9 AP respectively in the supervised setting of V3Det. For the open-vocabulary setting, Prova achieves a new state-of-the-art performance with 32.8 base AP and 11.0 novel AP, which is of 2.6 and 4.3 gain over the previous methods.

Large multimodal language models (MLLMs) have revolutionized natural language processing and visual understanding, but often contain outdated or inaccurate information. Current multimodal knowledge editing evaluations are limited in scope and potentially biased, focusing on narrow tasks and failing to assess the impact on in-domain samples. To address these issues, we introduce ComprehendEdit, a comprehensive benchmark comprising eight diverse tasks from multiple datasets. We propose two novel metrics: Knowledge Generalization Index (KGI) and Knowledge Preservation Index (KPI), which evaluate editing effects on in-domain samples without relying on AI-synthetic samples. Based on insights from our framework, we establish Hierarchical In-Context Editing (HICE), a baseline method employing a two-stage approach that balances performance across all metrics. This study provides a more comprehensive evaluation framework for multimodal knowledge editing, reveals unique challenges in this field, and offers a baseline method demonstrating improved performance. Our work opens new perspectives for future research and provides a foundation for developing more robust and effective editing techniques for MLLMs.

The field of music generation has seen a surge of interest from both academia and industry, with innovative platforms such as Suno, Udio, and SkyMusic earning widespread recognition. However, the challenge of music infilling—modifying specific music segments without reconstructing the entire piece—remains a significant hurdle for both audio-based and symbolic-based models, limiting their adaptability and practicality. In this paper, we address symbolic music infilling by introducing the Collaborative Music Inpainter (CMI), an advanced human-in-the-loop (HITL) model for music infilling. The CMI features the Joint Embedding Predictive Autoregressive Generative Architecture (JEP-AGA), which learns the high-level predictive representations of the masked part that needs to be infilled during the autoregressive generative process, akin to how humans perceive and interpret music. The newly developed Dynamic Interaction Learner (DIL) achieves HITL by iteratively refining the infilled output based on user interactions alone, significantly reducing the interaction cost without requiring further input. Experimental results confirm CMI’s superior performance in music infilling, demonstrating its efficiency in producing high-quality music.

Oracle Bone Inscriptions (OBIs), as the earliest systematically organized pictographic script in China, hold significant importance in the study of the origins of Chinese civilization. Of the approximately 4,500 excavated OBI characters, only about one-third have been deciphered, leaving the remaining characters shrouded in mystery. Over the past decade, an increasing number of researchers have attempted to leverage artificial intelligence to assist in deciphering OBIs, but these efforts have not yet fully met the demands of this challenging objective. In this paper, we identify a key task—Component-Level OBI Segmentation—based on a successful deciphering case from 2018. This task aims to help experts quickly identify specific components within OBIs, thereby accelerating the deciphering process. Accordingly, we propose a new model to accomplish this task. Our model leverages a small amount of annotated data and a large amount of weakly annotated data and incorporates expert-provided prior knowledge, i.e., stroke rules, to automatically segment OBI components. Additionally, we train a series of auxiliary classifiers to evaluate the segmentation results during the test stage. We also invite experts to conduct a professional assessment of the results, which we cross-validated against our proposed evaluation metrics. Experimental results demonstrate that our method can accurately and clearly present the segmented components to experts.

In this paper, we present a novel diffusion model-based monaural speech enhancement method. Our approach incorporates the separate estimation of speech spectra's magnitude and phase in two diffusion networks. Throughout the diffusion process, noise clips from real-world noise interferences are added gradually to the clean speech spectra and a noise-aware reverse process is proposed to learn how to generate both clean speech spectra and noise spectra. Furthermore, to fully leverage the intrinsic relationship between magnitude and phase, we introduce a complex-cycle-consistent (CCC) mechanism that uses the estimated magnitude to map the phase, and vice versa. We implement this algorithm within a phase-aware speech enhancement diffusion model (SEDM). We conduct extensive experiments on public datasets to demonstrate the effectiveness of our method, highlighting the significant benefits of exploiting the intrinsic relationship between phase and magnitude information to enhance speech. The comparison to conventional diffusion models demonstrates the superiority of SEDM.

This paper introduces a SAT-based technique that calculates a compact and complete symmetry-break for finite model finding, with the focus on structures with a single binary operation (magmas). Classes of algebraic structures are typically described as first-order logic formulas and the concrete algebras are models of these formulas. Such models include an enormous number of isomorphic, i.e. symmetric, algebras. A complete symmetry-break is a formula that has as models, exactly one canonical representative from each equivalence class of algebras. Thus, we enable answering questions about properties of the models so that computation and search are restricted to the set of canonical representations. For instance, we can answer the question: How many non-isomorphic semigroups are there of size n? Such questions can be answered by counting the satisfying assignments of a SAT formula, which already filters out non-isomorphic models. The introduced technique enables us calculating numbers of algebraic structures not present in the literature and going beyond the possibilities of pure enumeration approaches.

Artificial Intelligence (AI) has impacted the world tremendously in the last decade, causing an increased demand for accessible AI education globally. Students benefit from studying AI earlier in the curriculum; however, AI courses can require a range of prerequisites, which can be structured differently in various educational contexts. In this paper, we study the curriculum structure of AI, Machine Learning (ML), and Data Science (DS) courses in Canadian Universities and compare it with that of US Research-1 institutions. There are many similarities between AI, ML, and DS courses in Canada and the US. For example, DS courses tend to be more accessible earlier in the CS curriculum compared to AI and ML. However, there are key differences between the two countries, with Canadian AI, ML, and DS courses generally being a part of a longer prerequisites chain, and Canadian CS departments offering fewer DS courses. Still, both Canadian and US institutions find innovative ways to introduce AI earlier in the curriculum, including via interdisciplinary courses and specialized courses with few prerequisites. This study corroborates earlier work in recognizing diversity in curricular frameworks in North America and recommends curricular revisions and early academic advising to ensure access to AI courses.

Demonstration selection algorithms play a crucial role in optimizing Large Language Models' (LLMs) in-context learning performance. Despite numerous proposed algorithms, their comparative effectiveness remains understudied. We present a comprehensive evaluation of six state-of-the-art demonstration selection algorithms across five datasets, examining both their effectiveness and computational efficiency. Our findings reveal significant trade-offs: while some demonstration selection algorithms achieve superior accuracy, they incur substantial computational costs. We also discover that increasing demonstration examples doesn't consistently improve performance, and some sophisticated algorithms struggle to outperform random selection in certain scenarios. These insights provide valuable benchmarks for future algorithm development and practical implementation. Our code is available at https://github.com/Tizzzzy/Demonstration_Selection_Overview.

Graph unlearning technology has become increasingly important since the advent of the `right to be forgotten' and the growing concerns about the privacy and security of artificial intelligence. Graph unlearning aims to quickly eliminate the effects of specific data on graph neural networks (GNNs). However, most existing deterministic graph unlearning frameworks follow a balanced partition-submodel training-aggregation paradigm, resulting in a lack of structural information between subgraph neighborhoods and redundant unlearning parameter calculations. To address this issue, we propose a novel Graph Structure Mapping Unlearning paradigm (GSMU) and a novel method based on it named Community-centric Graph Eraser (CGE). CGE maps community subgraphs to nodes, thereby enabling the reconstruction of a node-level unlearning operation within a reduced mapped graph. CGE makes the exponential reduction of both the amount of training data and the number of unlearning parameters. Extensive experiments conducted on five real-world datasets and three widely used GNN backbones have verified the high performance and efficiency of our CGE method, highlighting its potential in the field of graph unlearning.

Variational autoencoder performs well in community detection on static networks, but it is difficult to directly extend to continuous dynamic networks. The main reason is that traditional methods mainly rely on adjacency structures to complete the inference and generation processes. However, continuous dynamic networks cannot be described by this structure because the inherent timeliness and causality information of the network would be lost. To address this issue, we propose a novel variational autoencoder, CT-VAE, for community detection in continuous dynamic networks, along with its scalable variant, CT-CAVAE. By conceptualizing node interactions as event streams and adopting the Hawkes process to capture temporal dynamics and causality, and incorporating them into the inference process, CT-VAE can effectively extend the traditional inference approach to continuous dynamic networks. Additionally, in the generation phase, CT-VAE combines pseudo-labeling and compact constraint strategies to facilitate the reconstruction process of non-adjacent structures. For the scalable variant, CT-CAVAE, end-to-end community detection is achieved by cleverly combining Gaussian mixture distribution. Extensive experimental results demonstrate that the proposed CT-VAE and CT-CAVAE achieve more favorable performance compared with the state-of-the-art baselines.

The increasing adoption of conversational agents powered by large language models (LLMs) raises questions about its effects across culturally diverse interactions. While these agents are linguistically versatile and multilingual, their ability to adapt along cultural dimensions--defined as geographically and communally nurtured sets of values and behavioral norms--lacks close scrutiny of both their design and deployment. To achieve inclusive conversational AI, it is essential to understand how agents adapt to users from diverse cultural backgrounds. In this study, we analyze dialogues between human users from different countries and LLM-powered agents to examine how both parties adapt their word use, a salient aspect of linguistic styles, toward one another throughout casual conversations. Our analysis reveals that LLMs exhibit varying degrees of style matching based on users' national cultures and demonstrate asymmetric adaptation when interacting with culturally diverse users. Moreover, we observe a reciprocal dynamic where both the LLMs and users from certain cultures adjust their styles in response to one another. Additionally, our findings support the hypothesis that LLMs and users naturally converge in conversational styles over the course of interactions, mirroring the dynamics of human conversations that accommodate and converge. To develop localized and culturally aware agents, there's a potential to utilize such cross-cultural convergence process during fine-tuning to align LLMs.