MouSi: Poly-Visual-Expert Vision-Language Models

Current large vision-language models (VLMs) often encounter challenges such as insufficient capabilities of a single visual component and excessively long visual tokens. These issues can limit the model's effectiveness in accurately interpreting complex visual information and over-lengthy contextual information. Addressing these challenges is crucial for enhancing the performance and applicability of VLMs. This paper proposes the use of ensemble experts technique to synergizes the capabilities of individual visual encoders, including those skilled in image-text matching, OCR, image segmentation, etc. This technique introduces a fusion network to unify the processing of outputs from different visual experts, while bridging the gap between image encoders and pre-trained LLMs. In addition, we explore different positional encoding schemes to alleviate the waste of positional encoding caused by lengthy image feature sequences, effectively addressing the issue of position overflow and length limitations. For instance, in our implementation, this technique significantly reduces the positional occupancy in models like SAM, from a substantial 4096 to a more efficient and manageable 64 or even down to 1. Experimental results demonstrate that VLMs with multiple experts exhibit consistently superior performance over isolated visual encoders and mark a significant performance boost as more experts are integrated. We have open-sourced the training code used in this report. All of these resources can be found on our project website.

From GPT-4 to Gemini and Beyond: Assessing the Landscape of MLLMs on Generalizability, Trustworthiness and Causality through Four Modalities

Multi-modal Large Language Models (MLLMs) have shown impressive abilities in generating reasonable responses with respect to multi-modal contents. However, there is still a wide gap between the performance of recent MLLM-based applications and the expectation of the broad public, even though the most powerful OpenAI's GPT-4 and Google's Gemini have been deployed. This paper strives to enhance understanding of the gap through the lens of a qualitative study on the generalizability, trustworthiness, and causal reasoning capabilities of recent proprietary and open-source MLLMs across four modalities: ie, text, code, image, and video, ultimately aiming to improve the transparency of MLLMs. We believe these properties are several representative factors that define the reliability of MLLMs, in supporting various downstream applications. To be specific, we evaluate the closed-source GPT-4 and Gemini and 6 open-source LLMs and MLLMs. Overall we evaluate 230 manually designed cases, where the qualitative results are then summarized into 12 scores (ie, 4 modalities times 3 properties). In total, we uncover 14 empirical findings that are useful to understand the capabilities and limitations of both proprietary and open-source MLLMs, towards more reliable downstream multi-modal applications.

GPS: Graph Contrastive Learning via Multi-scale Augmented Views from Adversarial Pooling

Self-supervised graph representation learning has recently shown considerable promise in a range of fields, including bioinformatics and social networks. A large number of graph contrastive learning approaches have shown promising performance for representation learning on graphs, which train models by maximizing agreement between original graphs and their augmented views (i.e., positive views). Unfortunately, these methods usually involve pre-defined augmentation strategies based on the knowledge of human experts. Moreover, these strategies may fail to generate challenging positive views to provide sufficient supervision signals. In this paper, we present a novel approach named Graph Pooling ContraSt (GPS) to address these issues. Motivated by the fact that graph pooling can adaptively coarsen the graph with the removal of redundancy, we rethink graph pooling and leverage it to automatically generate multi-scale positive views with varying emphasis on providing challenging positives and preserving semantics, i.e., strongly-augmented view and weakly-augmented view. Then, we incorporate both views into a joint contrastive learning framework with similarity learning and consistency learning, where our pooling module is adversarially trained with respect to the encoder for adversarial robustness. Experiments on twelve datasets on both graph classification and transfer learning tasks verify the superiority of the proposed method over its counterparts.

PolyCF: Towards the Optimal Spectral Graph Filters for Collaborative Filtering

Collaborative Filtering (CF) is a pivotal research area in recommender systems that capitalizes on collaborative similarities between users and items to provide personalized recommendations. With the remarkable achievements of node embedding-based Graph Neural Networks (GNNs), we explore the upper bounds of expressiveness inherent to embedding-based methodologies and tackle the challenges by reframing the CF task as a graph signal processing problem. To this end, we propose PolyCF, a flexible graph signal filter that leverages polynomial graph filters to process interaction signals. PolyCF exhibits the capability to capture spectral features across multiple eigenspaces through a series of Generalized Gram filters and is able to approximate the optimal polynomial response function for recovering missing interactions. A graph optimization objective and a pair-wise ranking objective are jointly used to optimize the parameters of the convolution kernel. Experiments on three widely adopted datasets demonstrate the superiority of PolyCF over current state-of-the-art CF methods. Moreover, comprehensive studies empirically validate each component's efficacy in the proposed PolyCF.

Preparing Lessons for Progressive Training on Language Models

The rapid progress of Transformers in artificial intelligence has come at the cost of increased resource consumption and greenhouse gas emissions due to growing model sizes. Prior work suggests using pretrained small models to improve training efficiency, but this approach may not be suitable for new model structures. On the other hand, training from scratch can be slow, and progressively stacking layers often fails to achieve significant acceleration. To address these challenges, we propose a novel method called Apollo, which prep\textbf{a}res lessons for ex\textbf{p}anding \textbf{o}perations by \textbf{l}earning high-\textbf{l}ayer functi\textbf{o}nality during training of low layers. Our approach involves low-value-prioritized sampling (LVPS) to train different depths and weight sharing to facilitate efficient expansion. We also introduce an interpolation method for stable model depth extension. Experiments demonstrate that Apollo achieves state-of-the-art acceleration ratios, even rivaling methods using pretrained models, making it a universal and efficient solution for training deep models while reducing time, financial, and environmental costs.

Reframing Tax Law Entailment as Analogical Reasoning

Statutory reasoning refers to the application of legislative provisions to a series of case facts described in natural language. We re-frame statutory reasoning as an analogy task, where each instance of the analogy task involves a combination of two instances of statutory reasoning. This increases the dataset size by two orders of magnitude, and introduces an element of interpretability. We show that this task is roughly as difficult to Natural Language Processing models as the original task. Finally, we come back to statutory reasoning, solving it with a combination of a retrieval mechanism and analogy models, and showing some progress on prior comparable work.

A Survey on Graph Neural Networks in Intelligent Transportation Systems

Intelligent Transportation System (ITS) is vital in improving traffic congestion, reducing traffic accidents, optimizing urban planning, etc. However, due to the complexity of the traffic network, traditional machine learning and statistical methods are relegated to the background. With the advent of the artificial intelligence era, many deep learning frameworks have made remarkable progress in various fields and are now considered effective methods in many areas. As a deep learning method, Graph Neural Networks (GNNs) have emerged as a highly competitive method in the ITS field since 2019 due to their strong ability to model graph-related problems. As a result, more and more scholars pay attention to the applications of GNNs in transportation domains, which have shown excellent performance. However, most of the research in this area is still concentrated on traffic forecasting, while other ITS domains, such as autonomous vehicles and urban planning, still require more attention. This paper aims to review the applications of GNNs in six representative and emerging ITS domains: traffic forecasting, autonomous vehicles, traffic signal control, transportation safety, demand prediction, and parking management. We have reviewed extensive graph-related studies from 2018 to 2023, summarized their methods, features, and contributions, and presented them in informative tables or lists. Finally, we have identified the challenges of applying GNNs to ITS and suggested potential future directions.

RJUA-QA: A Comprehensive QA Dataset for Urology

We introduce RJUA-QA, a novel medical dataset for question answering (QA) and reasoning with clinical evidence, contributing to bridge the gap between general large language models (LLMs) and medical-specific LLM applications. RJUA-QA is derived from realistic clinical scenarios and aims to facilitate LLMs in generating reliable diagnostic and advice. The dataset contains 2,132 curated Question-Context-Answer pairs, corresponding about 25,000 diagnostic records and clinical cases. The dataset covers 67 common urological disease categories, where the disease coverage exceeds 97.6\% of the population seeking medical services in urology. Each data instance in RJUA-QA comprises: (1) a question mirroring real patient to inquiry about clinical symptoms and medical conditions, (2) a context including comprehensive expert knowledge, serving as a reference for medical examination and diagnosis, (3) a doctor response offering the diagnostic conclusion and suggested examination guidance, (4) a diagnosed clinical disease as the recommended diagnostic outcome, and (5) clinical advice providing recommendations for medical examination. RJUA-QA is the first medical QA dataset for clinical reasoning over the patient inquiries, where expert-level knowledge and experience are required for yielding diagnostic conclusions and medical examination advice. A comprehensive evaluation is conducted to evaluate the performance of both medical-specific and general LLMs on the RJUA-QA dataset. Our data is are publicly available at \url{https://github.com/alipay/RJU_Ant_QA}.

A Survey of Reasoning with Foundation Models

Reasoning, a crucial ability for complex problem-solving, plays a pivotal role in various real-world settings such as negotiation, medical diagnosis, and criminal investigation. It serves as a fundamental methodology in the field of Artificial General Intelligence (AGI). With the ongoing development of foundation models, there is a growing interest in exploring their abilities in reasoning tasks. In this paper, we introduce seminal foundation models proposed or adaptable for reasoning, highlighting the latest advancements in various reasoning tasks, methods, and benchmarks. We then delve into the potential future directions behind the emergence of reasoning abilities within foundation models. We also discuss the relevance of multimodal learning, autonomous agents, and super alignment in the context of reasoning. By discussing these future research directions, we hope to inspire researchers in their exploration of this field, stimulate further advancements in reasoning with foundation models, and contribute to the development of AGI.

LLMEval: A Preliminary Study on How to Evaluate Large Language Models

Recently, the evaluation of Large Language Models has emerged as a popular area of research. The three crucial questions for LLM evaluation are ``what, where, and how to evaluate''. However, the existing research mainly focuses on the first two questions, which are basically what tasks to give the LLM during testing and what kind of knowledge it should deal with. As for the third question, which is about what standards to use, the types of evaluators, how to score, and how to rank, there hasn't been much discussion. In this paper, we analyze evaluation methods by comparing various criteria with both manual and automatic evaluation, utilizing onsite, crowd-sourcing, public annotators and GPT-4, with different scoring methods and ranking systems. We propose a new dataset, LLMEval and conduct evaluations on 20 LLMs. A total of 2,186 individuals participated, leading to the generation of 243,337 manual annotations and 57,511 automatic evaluation results. We perform comparisons and analyses of different settings and conduct 10 conclusions that can provide some insights for evaluating LLM in the future. The dataset and the results are publicly available at https://github.com/llmeval .