This study examines conversational business analytics, an approach that utilizes AI to address the technical competency gaps that hinder end users from effectively using traditional self-service analytics. By facilitating natural language interactions, conversational business analytics aims to empower end users to independently retrieve data and generate insights. The analysis focuses on Text-to-SQL as a representative technology for translating natural language requests into SQL statements. Developing theoretical models grounded in expected utility theory, this study identifies the conditions under which conversational business analytics, through partial or full support, can outperform delegation to human experts. The results indicate that partial support, focusing solely on information generation by AI, is viable when the accuracy of AI-generated SQL queries leads to a profit that surpasses the performance of a human expert. In contrast, full support includes not only information generation but also validation through explanations provided by the AI, and requires sufficiently high validation effectiveness to be reliable. However, user-based validation presents challenges, such as misjudgment and rejection of valid SQL queries, which may limit the effectiveness of conversational business analytics. These challenges underscore the need for robust validation mechanisms, including improved user support, automated processes, and methods for assessing quality independent of the technical competency of end users.

The advancement of Multimodal Large Language Models (MLLMs) has enabled significant progress in multimodal understanding, expanding their capacity to analyze video content. However, existing evaluation benchmarks for MLLMs primarily focus on abstract video comprehension, lacking a detailed assessment of their ability to understand video compositions, the nuanced interpretation of how visual elements combine and interact within highly compiled video contexts. We introduce VidComposition, a new benchmark specifically designed to evaluate the video composition understanding capabilities of MLLMs using carefully curated compiled videos and cinematic-level annotations. VidComposition includes 982 videos with 1706 multiple-choice questions, covering various compositional aspects such as camera movement, angle, shot size, narrative structure, character actions and emotions, etc. Our comprehensive evaluation of 33 open-source and proprietary MLLMs reveals a significant performance gap between human and model capabilities. This highlights the limitations of current MLLMs in understanding complex, compiled video compositions and offers insights into areas for further improvement. The leaderboard and evaluation code are available at https://yunlong10.github.io/VidComposition/.

Event Causality Identification (ECI) has become a crucial task in Natural Language Processing (NLP), aimed at automatically extracting causalities from textual data. In this survey, we systematically address the foundational principles, technical frameworks, and challenges of ECI, offering a comprehensive taxonomy to categorize and clarify current research methodologies, as well as a quantitative assessment of existing models. We first establish a conceptual framework for ECI, outlining key definitions, problem formulations, and evaluation standards. Our taxonomy classifies ECI methods according to the two primary tasks of sentence-level (SECI) and document-level (DECI) event causality identification. For SECI, we examine feature pattern-based matching, deep semantic encoding, causal knowledge pre-training and prompt-based fine-tuning, and external knowledge enhancement methods. For DECI, we highlight approaches focused on event graph reasoning and prompt-based techniques to address the complexity of cross-sentence causal inference. Additionally, we analyze the strengths, limitations, and open challenges of each approach. We further conduct an extensive quantitative evaluation of various ECI methods on two benchmark datasets. Finally, we explore future research directions, highlighting promising pathways to overcome current limitations and broaden ECI applications.

While small language models (SLMs) show promises for mobile deployment, their real-world performance and applications on smartphones remains underexplored. We present SlimLM, a series of SLMs optimized for document assistance tasks on mobile devices. Through extensive experiments on a Samsung Galaxy S24, we identify the optimal trade-offs between model size (ranging from 125M to 7B parameters), context length, and inference time for efficient on-device processing. SlimLM is pre-trained on SlimPajama-627B and fine-tuned on DocAssist, our constructed dataset for summarization, question answering and suggestion tasks. Our smallest model demonstrates efficient performance on S24, while larger variants offer enhanced capabilities within mobile constraints. We evaluate SlimLM against existing SLMs, showing comparable or superior performance and offering a benchmark for future research in on-device language models. We also provide an Android application, offering practical insights into SLM deployment. Our findings provide valuable insights and illuminate the capabilities of running advanced language models on high-end smartphones, potentially reducing server costs and enhancing privacy through on-device processing.

This paper explores a new black-box, zero-shot language model inversion problem and proposes an innovative framework for prompt reconstruction using only text outputs from a language model. Leveraging a large language model alongside an optimization algorithm, the proposed method effectively recovers prompts with minimal resources. Experimental results on several datasets derived from public sources indicate that the proposed approach achieves high-quality prompt recovery and generates prompts more similar to the originals than current state-of-the-art methods. Additionally, the use-case study demonstrates the method's strong potential for generating high-quality text data.

Thanks to unprecedented language understanding and generation capabilities of large language model (LLM), Retrieval-augmented Code Generation (RaCG) has recently been widely utilized among software developers. While this has increased productivity, there are still frequent instances of incorrect codes being provided. In particular, there are cases where plausible yet incorrect codes are generated for queries from users that cannot be answered with the given queries and API descriptions. This study proposes a task for evaluating answerability, which assesses whether valid answers can be generated based on users' queries and retrieved APIs in RaCG. Additionally, we build a benchmark dataset called Retrieval-augmented Code Generability Evaluation (RaCGEval) to evaluate the performance of models performing this task. Experimental results show that this task remains at a very challenging level, with baseline models exhibiting a low performance of 46.7%. Furthermore, this study discusses methods that could significantly improve performance. Figure 1: An example of an LLM generating plausible code even when the request is made outside the functionality provided by the library.

This paper presents a new approach to urban sustainability assessment through the use of Large Language Models (LLMs) to streamline the use of the ISO 37101 framework to automate and standardise the assessment of urban initiatives against the six "sustainability purposes" and twelve "issues" outlined in the standard. The methodology includes the development of a custom prompt based on the standard definitions and its application to two different datasets: 527 projects from the Paris Participatory Budget and 398 activities from the PROBONO Horizon 2020 project. The results show the effectiveness of LLMs in quickly and consistently categorising different urban initiatives according to sustainability criteria. The approach is particularly promising when it comes to breaking down silos in urban planning by providing a holistic view of the impact of projects. The paper discusses the advantages of this method over traditional human-led assessments, including significant time savings and improved consistency. However, it also points out the importance of human expertise in interpreting results and ethical considerations. This study hopefully can contribute to the growing body of work on AI applications in urban planning and provides a novel method for operationalising standardised sustainability frameworks in different urban contexts.

Large language models (LLMs) have been widely deployed as the backbone with additional tools and text information for real-world applications. However, integrating external information into LLM-integrated applications raises significant security concerns. Among these, prompt injection attacks are particularly threatening, where malicious instructions injected in the external text information can exploit LLMs to generate answers as the attackers desire. While both training-time and test-time defense methods have been developed to mitigate such attacks, the unaffordable training costs associated with training-time methods and the limited effectiveness of existing test-time methods make them impractical. This paper introduces a novel test-time defense strategy, named Formatting AuThentication with Hash-based tags (FATH). Unlike existing approaches that prevent LLMs from answering additional instructions in external text, our method implements an authentication system, requiring LLMs to answer all received instructions with a security policy and selectively filter out responses to user instructions as the final output. To achieve this, we utilize hash-based authentication tags to label each response, facilitating accurate identification of responses according to the user's instructions and improving the robustness against adaptive attacks. Comprehensive experiments demonstrate that our defense method can effectively defend against indirect prompt injection attacks, achieving state-of-the-art performance under Llama3 and GPT3.5 models across various attack methods. Our code is released at: https://github.com/Jayfeather1024/FATH

We present MMDS, a system capable of recognizing medical images and patient facial details, and providing professional medical diagnoses. The system consists of two core components:The first component is the analysis of medical images and videos. We trained a specialized multimodal medical model capable of interpreting medical images and accurately analyzing patients' facial emotions and facial paralysis conditions. The model achieved an accuracy of 72.59% on the FER2013 facial emotion recognition dataset, with a 91.1% accuracy in recognizing the "happy" emotion. In facial paralysis recognition, the model reached an accuracy of 92%, which is 30% higher than that of GPT-4o. Based on this model, we developed a parser for analyzing facial movement videos of patients with facial paralysis, achieving precise grading of the paralysis severity. In tests on 30 videos of facial paralysis patients, the system demonstrated a grading accuracy of 83.3%.The second component is the generation of professional medical responses. We employed a large language model, integrated with a medical knowledge base, to generate professional diagnoses based on the analysis of medical images or videos. The core innovation lies in our development of a department-specific knowledge base routing management mechanism, in which the large language model categorizes data by medical departments and, during the retrieval process, determines the appropriate knowledge base to query. This significantly improves retrieval accuracy in the RAG (retrieval-augmented generation) process.

Retrieval-Augmented Generation (RAG), while serving as a viable complement to large language models (LLMs), often overlooks the crucial aspect of text chunking within its pipeline, which impacts the quality of knowledge-intensive tasks. This paper introduces the concept of Meta-Chunking, which refers to a granularity between sentences and paragraphs, consisting of a collection of sentences within a paragraph that have deep linguistic logical connections. To implement Meta-Chunking, we designed Perplexity (PPL) Chunking, which balances performance and speed, and precisely identifies the boundaries of text chunks by analyzing the characteristics of context perplexity distribution. Additionally, considering the inherent complexity of different texts, we propose a strategy that combines PPL Chunking with dynamic merging to achieve a balance between fine-grained and coarse-grained text chunking. Experiments conducted on eleven datasets demonstrate that Meta-Chunking can more efficiently improve the performance of singlehop and multi-hop question answering based on RAG. For instance, on the 2Wiki-MultihopQA dataset, it outperforms similarity chunking by 1.32 while only consuming 45.8% of the time. Furthermore, through the analysis of models of various scales and types, we observed that PPL Chunking exhibits notable flexibility and adaptability. This is particularly relevant in knowledge-intensive tasks like open-domain question answering (Lazaridou et al., 2022). By integrating two key components: the retriever and the generator, this technology enables more precise responses to input queries (Singh et al., 2021; Lin et al., 2023). While the feasibility of the retrieval-augmentation strategy has been widely demonstrated through practice, its effectiveness heavily relies on the relevance and accuracy of the retrieved documents (Li et al., 2022; Tan et al., 2022). The introduction of excessive redundant or incomplete information through retrieval not only fails to enhance the performance of the generation model but may also lead to a decline in answer quality (Shi et al., 2023; Yan et al., 2024).