United States

Compositional relational reasoning (CRR) is a hallmark of human intelligence, but we lack a clear understanding of whether and how existing transformer large language models (LLMs) can solve CRR tasks. To enable systematic exploration of the CRR capability of LLMs, we first propose a new synthetic benchmark called Generalized Associative Recall (GAR) by integrating and generalizing the essence of several tasks in mechanistic interpretability (MI) study in a unified framework. Evaluation shows that GAR is challenging enough for existing LLMs, revealing their fundamental deficiency in CRR. Meanwhile, it is easy enough for systematic MI study. Then, to understand how LLMs solve GAR tasks, we use attribution patching to discover the core circuits reused by Vicuna-33B across different tasks, and a set of vital attention heads. Intervention experiments show that the correct functioning of these heads significantly impacts task performance. Especially, we identify two classes of heads whose activations represent the abstract notion of true and false in GAR tasks respectively. They play fundamental roles in CRR across various models and tasks.

AAAI 2025

Benchmarking and Understanding Compositional Relational Reasoning of LLMs

### Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

Message from the General Chair: 
*It is my great pleasure and honor to welcome you to the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025), held in beautiful Vienna, Austria, from July 27 to August 1, 2025. ACL2025continues our field’s tradition of excellence in scholarship, innovation, and inclusivity, and I am deeply grateful to the many volunteers who have worked tirelessly to bring this event to life.* 
[Read more](https://drive.google.com/file/d/1GI_hvOpjswAuYdUTromfeDiPpCcqidwg/view?usp=sharing)

To access this event page, you need to log in with the **email address you registered with**. Access credentials will be sent to your email from Underline - subject line "Welcome to ACL 2025". 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 need to log in with the **email address you registered with**. 

ACL 2025

Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

poster

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.



While pruning methods effectively maintain model perfor-
mance without extra training costs, they often focus solely
on preserving crucial connections, overlooking the impact
of pruned weights on subsequent fine-tuning or distillation,
leading to inefficiencies. Moreover, most compression tech-
niques for generative models have been developed primarily
for GANs, tailored to specific architectures like StyleGAN,
and research into compressing Diffusion models has just be-
gun. Even more, these methods are often applicable only to
GANs or Diffusion models, highlighting the need for ap-
proaches that work across both model types. In this paper, we introduce Singular Value Scaling (SVS), a versatile technique for refining pruned weights, applicable to both model types. Our analysis reveals that pruned weights often exhibit dominant singular vectors, hindering fine-tuning efficiency and leading to suboptimal performance compared to random initialization. Our method enhances weight initialization by minimizing the disparities between singular values of pruned
weights, thereby improving the fine-tuning process. This approach not only guides the compressed model toward superior solutions but also significantly speeds up fine-tuning. Extensive experiments on StyleGAN2, StyleGAN3 and DDPM demonstrate that SVS improves compression performance across model types without additional training costs.

Singular Value Scaling: Efficient Generative Model Compression via Pruned Weights Refinement

This paper studies decentralized optimization over a compact submanifold within a communication network of $n$ nodes, where each node possesses a smooth non-convex local cost function, and the goal is to jointly minimize the sum of these local costs. We focus particularly on the online setting, where local data is processed in real-time as it streams in, without the need for full data storage. We propose a decentralized projected Riemannian stochastic recursive momentum (DPRSRM) method that employs local hybrid stochastic gradient estimators and uses the network to track the global gradient. DPRSRM achieves an oracle complexity of  $\mathcal{O}(\epsilon^{-\frac{3}{2}})$, outperforming existing methods that have at most $\mathcal{O}(\epsilon^{-2})$ complexity. Our method requires only $\mathcal{O}(1)$ gradient evaluations per iteration for each local node and does not require restarting with a large batch gradient.   Furthermore, we demonstrate the effectiveness of our proposed methods compared to state-of-the-art ones through numerical experiments on principal component analysis problems and low-rank matrix completion.

Decentralized Projected Riemannian Stochastic Recursive Momentum Method for Nonconvex Optimization

While Nash equilibria are guaranteed to exist, they may exhibit dense support, making them difficult to understand and execute in some applications. In this paper, we study $k$-sparse commitments in games where one player is restricted to mixed strategies with support size at most $k$. Finding $k$-sparse commitments is known to be computationally hard. We start by showing several structural properties of $k$-sparse solutions, including that the optimal support may vary dramatically as $k$ increases. These results suggest that naive greedy or double-oracle-based approaches are unlikely to yield practical algorithms. We then develop a simple approach based on mixed integer linear programs (MILPs) for zero-sum games, general-sum Stackelberg games, and various forms of structured sparsity. We also propose practical algorithms for cases where one or both players have large (i.e., practically innumerable) action sets, utilizing a combination of MILPs and incremental strategy generation. We evaluate our methods on synthetic and real-world scenarios based on security applications. In both settings, we observe that even for small support sizes, we can obtain more than $90\%$ of the true Nash value while maintaining a reasonable runtime, demonstrating the significance of our formulation and algorithms.

Commitment to Sparse Strategies in Two-Player Games

REST APIs have become key components of web services. However, they often contain logic flaws resulting in server side errors or security vulnerabilities. HTTP requests are used as test cases to find and mitigate such issues. Existing methods to modify requests, including those using deep learning, suffer from limited performance and precision, relying on undirected search or making limited usage of the contextual information. In this paper we propose APIRL, a fully automated deep reinforcement learning tool for testing REST APIs. A key novelty of our approach is the use of feedback from a transformer moudle pre-trained on JSON-structured data, akin to that used in API responses. This allows APIRL to learn the subtleties relating to test outcomes, and generalise to unseen API endpoints. We show APIRL can find significantly more bugs than the state-of-the-art in real world REST APIs while minimising the number of required test cases. We also study how reward functions, and other key design choices, affect learnt policies in a thorough ablation study.

APIRL: Deep Reinforcement Learning for REST API Fuzzing

These recent years have witnessed that convolutional neural network (CNN)-based methods for detecting infrared small targets have achieved outstanding performance. However, these methods typically employ standard convolutions, neglecting to consider the spatial characteristics of the pixel distribution of infrared small targets. Therefore, we propose a novel Pinwheel-shaped Convolution (PConv) as a replacement for standard convolutions in the lower layers of the backbone network. PConv better aligns with the pixel gaussian spatial distribution of dim small targets, enhances feature extraction, significantly increases the receptive field, and introduces only a minimal increase in parameters.
Additionally, while recent loss functions combine scale and location losses, they do not adequately account for the varying sensitivity of these losses across different target scales, limiting detection performance on dim-small targets. To overcome this, we propose a Scale-based Dynamic (SD) Loss that dynamically adjusts the influence of scale and location losses based on target size, improving the network’s ability to detect targets of varying scales.
We construct a new benchmark, SIRST-UAVB, which is the largest and most challenging dataset to date for real-shot single-frame infrared small target detection. Lastly, by integrating PConv and SD Loss into the latest small target detection algorithms, we achieved significant performance improvements on IRSTD-1K and our SIRST-UAVB dataset, validating the effectiveness and generalizability of our approach.

Pinwheel-shaped Convolution and Scale-based Dynamic Loss for Infrared Small Target Detection

3D occupancy perception accurately estimates the volumetric status and semantic labels of a scene, attracting significant attention in the field of autonomous driving.
However, enhancing the model's ability to generalize across different driving scenarios or sensing systems, often requires redesigning the model or extra-expensive annotations.
To this end, following a comprehensive analysis of the occupancy model architecture, we proposed the \textbf{UGOCC} method that utilizes domain adaptation to efficiently harness unlabeled autonomous driving data, thereby enhancing the model’s generalizability.
Specifically, we design the depth fusion module by employing self-supervised depth estimation, and propose a strategy based on semantic attention and domain adversarial learning to improve the generalizability of the learnable fusion module. Additionally, we propose an OCC-specific pseudo-label selection tailored for semi-supervised learning, which optimizes the overall network's generalizability.
Our experiment results on two challenging datasets nuScenes and Waymo, demonstrate that our method not only achieves state-of-the-art  generalizability but also enhances the model's perceptual capabilities within the source domain by utilizing unlabeled data.

Enhancing Generalizability via Utilization of Unlabeled Data for Occupancy Perception

Most existing just noticeable difference (JND) methods primarily integrate specific masking effects in a single domain. However, these single-domain JND methods struggle with the structural discrepancies in multi-source content images, limiting their effectiveness in visual redundancy estimation. To address this  issue, we propose a dual domain encoder that combines spatial and frequency features to comprehensively capture visual patterns. Our design includes spatial pattern balance and frequency detail correction modules to balance global and local patterns and correct low- and high-frequency  distributions. Additionally, we develop a dual domain decoder to effectively extract multi-scale pattern redundancies and integrate them with detail redundancies in the frequency domain. Experiments demonstrate the effectiveness and robustness of our proposed method in handling structural discrepancies in multi-source content images.

DDJND: Dual Domain Just Noticeable Difference in Multi-Source Content Images with Structural Discrepancy

Anomaly detection on attributed networks has applications in various domains such as finance and email spam detection, thus gaining substantial attention. However, the current mainstream anomaly detection methods are mainly designed for centralized architectures, which may bring the risk of privacy leakage in some sensitive scenarios. As an emerging technology, federated graph learning aims to realize model training of graph data through distributed collaboration, while preserving data privacy. Nevertheless, when graph data is scattered across multiple clients, federated graph learning may not manage to harness the full potential of distributed data, yielding suboptimal performance. Taking into account this insight, we propose a federated graph anomaly detection framework utilizing contrastive self-supervised learning (FedCLGN). First, we put forward an augmentation method to maintain global negative pairs on the server side. This includes initial identification of anomaly nodes using pseudo-labels and extraction of node neighbor embedding data from clients. Then, we adopt graph diffusion to enhance the feature representation of nodes, capturing the global structure and local connection patterns. This strategy can strengthen the differentiation between positive and negative instance pairs. Finally, the effectiveness of our approach is verified by experimental results on four real graph datasets.

Federated Graph Anomaly Detection Through Contrastive Learning with Global Negative Pairs

Recent advancements in Large Language Models (LLMs) have markedly enhanced the interpretation and processing of tabular data, introducing previously unimaginable capabilities. Despite these achievements, LLMs still encounter significant challenges when applied in industrial scenarios, particularly due to the increased complexity of reasoning required with real-world tabular data, underscoring a notable disparity between academic benchmarks and practical applications. To address this discrepancy, we conduct a detailed investigation into the application of tabular data in industrial scenarios and propose a comprehensive and complex benchmark TableBench, including 18 fields within four major categories of table question answering (TableQA) capabilities. Furthermore, we introduce TableLLM, trained on our meticulously constructed training set TableInstruct, achieving comparable performance with GPT-3.5. Massive experiments conducted on TableBench indicate that both open-source and proprietary LLMs still have significant room for improvement to meet real-world demands, where the most advanced model, GPT-4, achieves only a modest score compared to humans.

TableBench: A Comprehensive and Complex Benchmark for Table Question Answering

Despite the advancements in large language models (LLMs) for mathematical reasoning, solving competition-level math problems remains a significant challenge, especially for open-source LLMs without external tools. We introduce the MMIQC dataset, comprising a mixture of processed web data and synthetic question-response pairs, aimed at enhancing the mathematical reasoning capabilities of base language models. Models fine-tuned on MMIQC consistently surpass their counterparts in performance on the MATH benchmark across various model sizes. Notably, Qwen-72B-MMIQC achieves a 45.0% accuracy, exceeding the previous open-source state-of-the-art by 8.2% and outperforming the initial version GPT-4 released in 2023. Extensive evaluation results on Hungarian high school finals suggest that such improvement can generalize to unseen data. Our ablation study on MMIQC reveals that a large part of the improvement can be attributed to our novel augmentation method, Iterative Question Composing (IQC), which involves iteratively composing new questions from seed problems using an LLM and applying rejection sampling through another LLM.

Downloads

Next from AAAI 2025

Singular Value Scaling: Efficient Generative Model Compression via Pruned Weights Refinement

Similar lecture

CA-GAR: Context-Aware Alignment of LLM Generation for Document Retrieval

.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 2025

Singular Value Scaling: Efficient Generative Model Compression via Pruned Weights Refinement

Similar lecture

CA-GAR: Context-Aware Alignment of LLM Generation for Document Retrieval

Downloads