Complex video question-answering (VQA) requires in-depth understanding of video contents including object and action recognition as well as video classification and summarization, which exhibits great potential in emerging applications in education and entertainment, etc. Multimodal large language models (MLLMs) may accomplish this task by grasping the intention of a question and decomposing it to a series of visual recognition sub-tasks to find out the answer with the help of an agent. To tackle this task, we first collect a new dedicated Complex VQA dataset named CVQA and then propose VQAGuider, an innovative framework planning a few atomic visual recognition tools by video-related API matching. VQAGuider facilitates a deep engagement with video content and precise responses to complex video-related questions by MLLMs, which is beyond aligning visual and language features for simple VQA tasks. Our experiments demonstrate VQAGuider is capable of navigating the complex VQA tasks by MLLMs and improves the accuracy by 29.6% and 17.2% on CVQA and the existing VQA datasets, respectively, highlighting its potential in advancing MLLMs’s capabilities in video understanding.

Large language models (LLMs) have demonstrated impressive performance on reasoning tasks, including mathematical reasoning. However, the current evaluation mostly focuses on carefully constructed benchmarks and neglects the consideration of real-world reasoning problems that present missing or contradictory conditions, known as ill-defined problems. To further study this problem, we develop a large-scale benchmark called Problems with Missing and Contradictory conditions (PMC) containing over 5,000 validated ill-defined mathematical problems. Our preliminary experiments through PMC reveal two challenges about existing methods: (1) traditional methods exhibit a trade-off between solving accuracy and rejection capabilities, and (2) formal methods struggle with modeling complex problems. To address these challenges, We develop Variable-Constraint Search (VCSearch), a training-free framework that leverages formal language to detect ill-defined problems, where a variable-constraint pair search strategy is incorporated to improve the modeling capability of formal language. Extensive experiments demonstrate that VCSearch improves the accuracy of identifying unsolvable problems by at least 12% across different LLMs, thus achieving stronger robust mathematical reasoning ability.

The deployment of Large Language Models (LLMs) faces significant challenges due to high computational costs,driving the demand for effective pruning techniques. Existing structured pruning methods employ uniform compression rates across network layers, neglecting the varying importance of different network depths. To address this limitation, we propose a novel optimization framework that directly minimizes global capability loss through layer-adaptive pruning rates. The framework formulates the pruning task as a combinatorial optimization problem constrained by a total parameter budget, and an efficient dynamic programming solution is derived to determine optimal layer-wise compression rates.Experiments demonstrate that, when tuning is not included, our approach achieves comparable performance with state-of-the-art methods at high pruning rates (37-50% reduction), and shows significant advantages at low pruning rates (13-25% reduction). When tuning is included, our method achieves the best performance among the compared methods.

In this article, we tackle the math word problem, namely, automatically answering a mathematical problem according to its textual description. Although recent methods have demonstrated their promising results, most of these methods are based on template-based generation scheme which results in limited generalization capability. To this end, we propose a novel human-like analogical learning method in a recall and learn manner. Our proposed framework is composed of modules of memory, representation, analogy, and reasoning, which are designed to make a new exercise by referring to the exercises learned in the past. Specifically, given a math word problem, the model first retrieves similar questions by a memory module and then encodes the unsolved problem and each retrieved question using a representation module. Moreover, to solve the problem in a way of analogy, an analogy module and a reasoning module with a copy mechanism are proposed to model the interrelationship between the problem and each retrieved question. Extensive experiments on two well-known datasets show the superiority of our proposed algorithm as compared to other state-of-the-art competitors from both overall performance comparison and micro-scope studies.