Recent advancements in multimodal large language models have successfully extended the Reason-Then-Respond paradigm to image-based reasoning, yet video-based reasoning remains an underdeveloped frontier, primarily due to the scarcity of high-quality reasoning-oriented data and effective training methodologies. To bridge this gap, we introduce DarkEventInfer and MixVidQA, two novel datasets specifically designed to stimulate the model's advanced video understanding and reasoning abilities. DarkEventinfer presents videos with masked event segments, requiring models to infer the obscured content based on contextual video cues. MixVidQA, on the other hand, presents interleaved video sequences composed of two distinct clips, challenging models to isolate and reason about one while disregarding the other. Leveraging these carefully curated training samples together with reinforcement learning guided by diverse reward functions, we develop VersaVid-R1, the first versatile video understanding and reasoning model under the Reason-Then-Respond paradigm capable of handling multiple-choice and open-ended question answering, as well as video captioning tasks. Extensive experiments demonstrate that VersaVid-R1 significantly outperforms existing models across a broad spectrum of benchmarks, covering video general understanding, cognitive reasoning, and captioning tasks.
The advent of multimodal large language models (MLLMs) has sparked interest in their application to electrocardiogram (ECG) analysis. However, existing ECG-focused MLLMs primarily focus on report generation tasks, often limited to single 12-lead, short-duration (10s) ECG inputs, thereby underutilizing the potential of MLLMs. To this end, we aim to develop a MLLM for ECG analysis that supports a broader range of tasks and more flexible ECG inputs. However, existing ECG-QA datasets are often monotonous. To address this gap, we first constructed the anyECG dataset, which encompasses a wide variety of tasks, including report generation, abnormal waveform localization, and open-ended question answering. In addition to standard hospital ECGs, we introduced long-duration reduced-lead ECGs for home environments and multiple ECG comparison scenarios commonly encountered in clinical practice. Furthermore, we propose the anyECG-chat model, which supports dynamic-length ECG inputs and multiple ECG inputs. We trained the model using a three-stage curriculum training recipe with the anyECG dataset. A comprehensive evaluation was conducted, demonstrating that anyECG-chat is capable of supporting various practical application scenarios, including not only common report generation tasks but also abnormal waveform localization for long-duration reduced-lead ECGs in home environments and comprehensive comparative analysis of multiple ECGs.
This research-to-practice work-in-progress (WIP) paper presents an AI-enabled smart tutor designed to provide homework assessment and feedback for students in an undergraduate circuit analysis course. We detail the tutor's design philosophy and core components, including open-ended question answering and homework feedback generation. The prompts are carefully crafted to optimize responses across different problems. The smart tutor was deployed on the Microsoft Azure platform and is currently in use in an undergraduate circuit analysis course at the School of Electrical and Computer Engineering in a large, public, research-intensive institution in the Southeastern United States. Beyond offering personalized instruction and feedback, the tutor collects student interaction data, which is summarized and shared with the course instructor. To evaluate its effectiveness, we collected student feedback, with 90.9% of responses indicating satisfaction with the tutor. Additionally, we analyze a subset of collected data on preliminary circuit analysis topics to assess tutor usage frequency for each problem and identify frequently asked questions. These insights help instructors gain real-time awareness of student difficulties, enabling more targeted classroom instruction. In future work, we will release a full analysis once the complete dataset is available after the Spring 2025 semester. We also explore the potential applications of this smart tutor across a broader range of engineering disciplines by developing improved prompts, diagram-recognition methods, and database management strategies, which remain ongoing areas of research.
Large Language Models (LLMs) have demonstrated remarkable capabilities across various tasks, yet they often exhibit a specific cultural biases, neglecting the values and linguistic diversity of low-resource regions. This cultural bias not only undermines universal equality, but also risks reinforcing stereotypes and perpetuating discrimination. To address this, we propose CulFiT, a novel culturally-aware training paradigm that leverages multilingual data and fine-grained reward modeling to enhance cultural sensitivity and inclusivity. Our approach synthesizes diverse cultural-related questions, constructs critique data in culturally relevant languages, and employs fine-grained rewards to decompose cultural texts into verifiable knowledge units for interpretable evaluation. We also introduce GlobalCultureQA, a multilingual open-ended question-answering dataset designed to evaluate culturally-aware responses in a global context. Extensive experiments on three existing benchmarks and our GlobalCultureQA demonstrate that CulFiT achieves state-of-the-art open-source model performance in cultural alignment and general reasoning.
Large Language Models (LLMs), despite their advancements, are fundamentally limited by their static parametric knowledge, hindering performance on tasks requiring open-domain up-to-date information. While enabling LLMs to interact with external knowledge environments is a promising solution, current efforts primarily address closed-end problems. Open-ended questions, which characterized by lacking a standard answer or providing non-unique and diverse answers, remain underexplored. To bridge this gap, we present O$^2$-Searcher, a novel search agent leveraging reinforcement learning to effectively tackle both open-ended and closed-ended questions in the open domain. O$^2$-Searcher leverages an efficient, locally simulated search environment for dynamic knowledge acquisition, effectively decoupling the external world knowledge from model's sophisticated reasoning processes. It employs a unified training mechanism with meticulously designed reward functions, enabling the agent to identify problem types and adapt different answer generation strategies. Furthermore, to evaluate performance on complex open-ended tasks, we construct O$^2$-QA, a high-quality benchmark featuring 300 manually curated, multi-domain open-ended questions with associated web page caches. Extensive experiments show that O$^2$-Searcher, using only a 3B model, significantly surpasses leading LLM agents on O$^2$-QA. It also achieves SOTA results on various closed-ended QA benchmarks against similarly-sized models, while performing on par with much larger ones.
Remote-sensing applications often run on edge hardware that cannot host today's 7B-parameter multimodal language models. This paper introduces TinyRS, the first 2B-parameter multimodal small language model (MSLM) optimized for remote sensing tasks, and TinyRS-R1, its reasoning-augmented variant. Built upon Qwen2-VL-2B, TinyRS is trained through a four-stage pipeline: pre-training on million satellite images, instruction tuning on visual instruction examples, fine-tuning with Chain-of-Thought (CoT) annotations from the proposed reasoning dataset, and alignment via Group Relative Policy Optimization (GRPO). TinyRS-R1 achieves or surpasses the performance of recent 7B-parameter remote sensing models across classification, VQA, visual grounding, and open-ended question answering-while requiring just one-third of the memory and latency. Our analysis shows that CoT reasoning substantially benefits spatial grounding and scene understanding, while the non-reasoning TinyRS excels in concise, latency-sensitive VQA tasks. TinyRS-R1 represents the first domain-specialized MSLM with GRPO-aligned CoT reasoning for general-purpose remote sensing.
Majority voting is considered an effective method to enhance chain-of-thought reasoning, as it selects the answer with the highest "self-consistency" among different reasoning paths (Wang et al., 2023). However, previous chain-of-thought reasoning methods typically generate only a single answer in each trial, thereby ignoring the possibility of other potential answers. As a result, these alternative answers are often overlooked in subsequent voting processes. In this work, we propose to generate ranked answers in each reasoning process and conduct ranked voting among multiple ranked answers from different responses, thereby making the overall self-consistency more reliable. Specifically, we use three ranked voting methods: Instant-runoff voting, Borda count voting, and mean reciprocal rank voting. We validate our methods on six datasets, including three multiple-choice and three open-ended question-answering tasks, using both advanced open-source and closed-source large language models. Extensive experimental results indicate that our proposed method outperforms the baselines, showcasing the potential of leveraging the information of ranked answers and using ranked voting to improve reasoning performance. The code is available at https://github.com/szu-tera/RankedVotingSC.
Recent research looks to harness the general knowledge and reasoning of large language models (LLMs) into agents that accomplish user-specified goals in interactive environments. Vision-language models (VLMs) extend LLMs to multi-modal data and provide agents with the visual reasoning necessary for new applications in areas such as computer automation. However, agent tasks emphasize skills where accessible open-weight VLMs lag behind their LLM equivalents. For example, VLMs are less capable of following an environment's strict output syntax requirements and are more focused on open-ended question answering. Overcoming these limitations requires supervised fine-tuning (SFT) on task-specific expert demonstrations. Our work approaches these challenges from an offline-to-online reinforcement learning (RL) perspective. RL lets us fine-tune VLMs to agent tasks while learning from the unsuccessful decisions of our own model or more capable (larger) models. We explore an off-policy RL solution that retains the stability and simplicity of the widely used SFT workflow while allowing our agent to self-improve and learn from low-quality datasets. We demonstrate this technique with two open-weight VLMs across three multi-modal agent domains.
The proliferation of large language models (LLMs) has significantly advanced information retrieval systems, particularly in response generation (RG). Unfortunately, LLMs often face knowledge conflicts between internal memory and retrievaled external information, arising from misinformation, biases, or outdated knowledge. These conflicts undermine response reliability and introduce uncertainty in decision-making. In this work, we analyze how LLMs navigate knowledge conflicts from an information-theoretic perspective and reveal that when conflicting and supplementary information exhibit significant differences, LLMs confidently resolve their preferences. However, when the distinction is ambiguous, LLMs experience heightened uncertainty. Based on this insight, we propose Swin-VIB, a novel framework that integrates a pipeline of variational information bottleneck models into adaptive augmentation of retrieved information and guiding LLM preference in response generation. Extensive experiments on single-choice, open-ended question-answering (QA), and retrieval augmented generation (RAG) validate our theoretical findings and demonstrate the efficacy of Swin-VIB. Notably, our method improves single-choice task accuracy by at least 7.54\% over competitive baselines.
Time series data are foundational in finance, healthcare, and energy domains. However, most existing methods and datasets remain focused on a narrow spectrum of tasks, such as forecasting or anomaly detection. To bridge this gap, we introduce Time Series Multi-Task Question Answering (Time-MQA), a unified framework that enables natural language queries across multiple time series tasks - numerical analytical tasks and open-ended question answering with reasoning. Central to Time-MQA is the TSQA dataset, a large-scale dataset containing $\sim$200k question-answer pairs derived from diverse time series spanning environment, traffic, etc. This comprehensive resource covers various time series lengths and promotes robust model development. We further demonstrate how continually pre-training large language models (Mistral 7B, Llama-3 8B, and Qwen-2.5 7B) on the TSQA dataset enhanced time series reasoning capabilities, moving beyond mere numeric tasks and enabling more advanced and intuitive interactions with temporal data. The complete TSQA dataset, models, executable codes, user study questionnaires for evaluation, and results have all been open-sourced.
In autonomous driving, open-ended question answering often suffers from unreliable evaluations because freeform responses require either complex metrics or subjective human judgment. To address this challenge, we introduce AutoDrive-QA, an automatic pipeline that converts existing driving QA datasets (including DriveLM, NuScenes-QA, and LingoQA) into a structured multiple-choice question (MCQ) format. This benchmark systematically assesses perception, prediction, and planning tasks, providing a standardized and objective evaluation framework. AutoDrive-QA employs an automated pipeline that leverages large language models (LLMs) to generate high-quality, contextually relevant distractors based on domain-specific error patterns commonly found in autonomous driving scenarios. To evaluate both general capabilities and generalization performance, we test the benchmark on three public datasets and conduct zero-shot experiments on an unseen dataset. The zero-shot evaluations reveal that GPT-4V leads with 69.57% accuracy -- achieving 74.94% in Perception, 65.33% in Prediction, and 68.45% in Planning -- demonstrating that while all models excel in Perception, they struggle in Prediction. Consequently, AutoDrive-QA establishes a rigorous, unbiased standard for integrating and evaluating different vision-language models across various autonomous driving datasets, thereby improving generalization in this field. We release all the codes in the AutoDrive-QA GitHub Repository.
Effective personalization of LLMs is critical for a broad range of user-interfacing applications such as virtual assistants and content curation. Inspired by the strong in-context learning capabilities of LLMs, we propose Few-Shot Preference Optimization (FSPO), which reframes reward modeling as a meta-learning problem. Under this framework, an LLM learns to quickly adapt to a user via a few labeled preferences from that user, constructing a personalized reward function for them. Additionally, since real-world preference data is scarce and challenging to collect at scale, we propose careful design choices to construct synthetic preference datasets for personalization, generating over 1M synthetic personalized preferences using publicly available LLMs. In particular, to successfully transfer from synthetic data to real users, we find it crucial for the data to exhibit both high diversity and coherent, self-consistent structure. We evaluate FSPO on personalized open-ended generation for up to 1,500 synthetic users across across three domains: movie reviews, pedagogical adaptation based on educational background, and general question answering, along with a controlled human study. Overall, FSPO achieves an 87% Alpaca Eval winrate on average in generating responses that are personalized to synthetic users and a 72% winrate with real human users in open-ended question answering.
The rapid development of large language models (LLMs) is redefining the landscape of human-computer interaction, and their integration into various user-service applications is becoming increasingly prevalent. However, transmitting user data to cloud-based LLMs presents significant risks of data breaches and unauthorized access to personal identification information. In this paper, we propose a privacy preservation pipeline for protecting privacy and sensitive information during interactions between users and LLMs in practical LLM usage scenarios. We construct SensitiveQA, the first privacy open-ended question-answering dataset. It comprises 57k interactions in Chinese and English, encompassing a diverse range of user-sensitive information within the conversations. Our proposed solution employs a multi-stage strategy aimed at preemptively securing user information while simultaneously preserving the response quality of cloud-based LLMs. Experimental validation underscores our method's efficacy in balancing privacy protection with maintaining robust interaction quality. The code and dataset are available at https://github.com/ligw1998/PRIV-QA.
We introduce Neptune, a benchmark for long video understanding that requires reasoning over long time horizons and across different modalities. Many existing video datasets and models are focused on short clips (10s-30s). While some long video datasets do exist, they can often be solved by powerful image models applied per frame (and often to very few frames) in a video, and are usually manually annotated at high cost. In order to mitigate both these problems, we propose a scalable dataset creation pipeline which leverages large models (VLMs and LLMs), to automatically generate dense, time-aligned video captions, as well as tough question answer decoy sets for video segments (up to 15 minutes in length). Our dataset Neptune covers a broad range of long video reasoning abilities and consists of a subset that emphasizes multimodal reasoning. Since existing metrics for open-ended question answering are either rule-based or may rely on proprietary models, we provide a new open source model-based metric GEM to score open-ended responses on Neptune. Benchmark evaluations reveal that most current open-source long video models perform poorly on Neptune, particularly on questions testing temporal ordering, counting and state changes. Through Neptune, we aim to spur the development of more advanced models capable of understanding long videos. The dataset is available at https://github.com/google-deepmind/neptune
Large language models have demonstrated impressive performance when integrated with vision models even enabling video understanding. However, evaluating these video models presents its own unique challenges, for which several benchmarks have been proposed. In this paper, we show that the currently most used video-language benchmarks can be solved without requiring much temporal reasoning. We identified three main issues in existing datasets: (i) static information from single frames is often sufficient to solve the tasks (ii) the text of the questions and candidate answers is overly informative, allowing models to answer correctly without relying on any visual input (iii) world knowledge alone can answer many of the questions, making the benchmarks a test of knowledge replication rather than visual reasoning. In addition, we found that open-ended question-answering benchmarks for video understanding suffer from similar issues while the automatic evaluation process with LLMs is unreliable, making it an unsuitable alternative. As a solution, we propose TVBench, a novel open-source video multiple-choice question-answering benchmark, and demonstrate through extensive evaluations that it requires a high level of temporal understanding. Surprisingly, we find that most recent state-of-the-art video-language models perform similarly to random performance on TVBench, with only a few models such as Qwen2-VL, and Tarsier clearly surpassing this baseline.
Recent works integrating Knowledge Graphs (KGs) have led to promising improvements in enhancing reasoning accuracy of Large Language Models (LLMs). However, current benchmarks focus mainly on closed-ended tasks, leaving a gap in the assessment of more complex real-world scenarios. This gap has also obscured the evaluation of KGs' potential to mitigate the problem of hallucination in LLMs. To fill the gap, we introduce OKGQA, a new benchmark specifically designed to assess LLMs enhanced with KGs under open-ended, real-world question answering scenarios. OKGQA is designed to closely reflect the complexities of practical applications using questions from different types, and incorporates specific metrics to measure both the reduction in hallucinations and the enhancement in reasoning capabilities. To consider the scenario in which KGs may have varying levels of mistakes, we propose another benchmark variant OKGQA-P to assess model performance when the semantics and structure of KGs are deliberately perturbed and contaminated. OKGQA aims to (1) explore whether KGs can make LLMs more trustworthy in an open-ended setting, and (2) conduct a comparative analysis to shed light on methods and future directions for leveraging KGs to reduce LLMs' hallucination. We believe that this study can facilitate a more complete performance comparison and encourage continuous improvement in integrating KGs with LLMs.
Refusal-Aware Instruction Tuning (RAIT) enables Large Language Models (LLMs) to refuse to answer unknown questions. By modifying responses of unknown questions in the training data to refusal responses such as "I don't know", RAIT enhances the reliability of LLMs and reduces their hallucination. Generally, RAIT modifies training samples based on the correctness of the initial LLM's response. However, this crude approach can cause LLMs to excessively refuse answering questions they could have correctly answered, the problem we call over-refusal. In this paper, we explore two primary causes of over-refusal: Static conflict occurs when similar samples within the LLM's feature space receive differing supervision signals (original vs. modified "I don't know"). Dynamic conflict arises as the LLM's evolving knowledge during SFT enables it to answer previously unanswerable questions, but the now-answerable training samples still retain the original "I don't know" supervision signals from the initial LLM state, leading to inconsistencies. These conflicts cause the trained LLM to misclassify known questions as unknown, resulting in over-refusal. To address this issue, we introduce Certainty Represented Knowledge Flow for Refusal-Aware Instructions Tuning (CRaFT). CRaFT centers on two main contributions: First, we additionally incorporate response certainty to selectively filter and modify data, reducing static conflicts. Second, we implement preliminary rehearsal training to characterize changes in the LLM's knowledge state, which helps mitigate dynamic conflicts during the fine-tuning process. We conducted extensive experiments on open-ended question answering and multiple-choice question task. Experiment results show that CRaFT can improve LLM's overall performance during the RAIT process. Code and data will be released at https://github.com/opendatalab/CRaFT .
The development of video large multimodal models (LMMs) has been hindered by the difficulty of curating large amounts of high-quality raw data from the web. To address this, we propose an alternative approach by creating a high-quality synthetic dataset specifically for video instruction-following, namely LLaVA-Video-178K. This dataset includes key tasks such as detailed captioning, open-ended question-answering (QA), and multiple-choice QA. By training on this dataset, in combination with existing visual instruction tuning data, we introduce LLaVA-Video, a new video LMM. Our experiments demonstrate that LLaVA-Video achieves strong performance across various video benchmarks, highlighting the effectiveness of our dataset. We plan to release the dataset, its generation pipeline, and the model checkpoints.
Large Language Models (LLMs) are the cornerstones of modern artificial intelligence systems. This paper introduces Juhaina, a Arabic-English bilingual LLM specifically designed to align with the values and preferences of Arabic speakers. Juhaina inherently supports advanced functionalities such as instruction following, open-ended question answering, information provisioning, and text processing. Our model contains 9.24 billion parameters and is trained on a context window of up to 8,192 tokens. This paper details the creation process of Juhaina and provides an extensive empirical evaluation. Furthermore, we identify the limitations of widely-adopted Open Arabic LLM Leaderboard (OALL) and propose a new evaluation benchmark, CamelEval. Our findings demonstrate that Juhaina surpasses existing LLMs of comparable sizes, such as the Llama and Gemma families, in generating helpful responses in Arabic, providing factually accurate information about the region, and understanding nuanced cultural aspects. We aspire for Juhaina to democratize cutting-edge AI technologies, serving over 400 million Arabic speakers by offering LLMs that not only communicate in their language but also comprehend their culture. We publicly release all models on Huggingface \url{https://huggingface.co/elmrc}.
Evaluating the output of generative large language models (LLMs) is challenging and difficult to scale. Many evaluations of LLMs focus on tasks such as single-choice question-answering or text classification. These tasks are not suitable for assessing open-ended question-answering capabilities, which are critical in domains where expertise is required. One such domain is health, where misleading or incorrect answers can have a negative impact on a user's well-being. Using human experts to evaluate the quality of LLM answers is generally considered the gold standard, but expert annotation is costly and slow. We present a method for evaluating LLM answers that uses ranking models trained on annotated document collections as a substitute for explicit relevance judgements and apply it to the CLEF 2021 eHealth dataset. In a user study, our method correlates with the preferences of a human expert (Kendall's $\tau=0.64$). It is also consistent with previous findings in that the quality of generated answers improves with the size of the model and more sophisticated prompting strategies.