Multi-page visual documents such as manuals, brochures, presentations, and posters convey key information through layout, colors, icons, and cross-slide references.While multimodal large language models (MLLMs) offer opportunities in document understanding, current systems struggle with complex, multi-page visual documents, particularly in fine-grained reasoning over elements and pages.We introduce SlideAgent, a versatile agentic framework for understanding multi-modal, multipage, and multi-layout documents, especially slide decks.SlideAgent employs specialized agents and decomposes reasoning into three specialized levels-global, page, and element-to construct a structured, query-agnostic representation that captures both overarching themes and detailed visual or textual cues.During inference, SlideAgent selectively activates specialized agents for multi-level reasoning and integrates their outputs into coherent, contextaware answers.Extensive experiments show that SlideAgent significantly improves accuracy over both proprietary (+7.9%) and opensource models (+9.8%).

Recent multimodal LLMs have shown promise in chart-based visual question answering, but their performance declines sharply on unannotated charts-those requiring precise visual interpretation rather than relying on textual shortcuts. To address this, we introduce ChartAgent, a novel agentic framework that explicitly performs visual reasoning directly within the chart's spatial domain. Unlike textual chain-of-thought reasoning, ChartAgent iteratively decomposes queries into visual subtasks and actively manipulates and interacts with chart images through specialized actions such as drawing annotations, cropping regions (e.g., segmenting pie slices, isolating bars), and localizing axes, using a library of chart-specific vision tools to fulfill each subtask. This iterative reasoning process closely mirrors human cognitive strategies for chart comprehension. ChartAgent achieves state-of-the-art accuracy on the ChartBench and ChartX benchmarks, surpassing prior methods by up to 16.07% absolute gain overall and 17.31% on unannotated, numerically intensive queries. Furthermore, our analyses show that ChartAgent is (a) effective across diverse chart types, (b) achieves the highest scores across varying visual and reasoning complexity levels, and (c) serves as a plug-and-play framework that boosts performance across diverse underlying LLMs. Our work is among the first to demonstrate visually grounded reasoning for chart understanding using tool-augmented multimodal agents.

Recent advancements in language modeling have shown promising results when applied to time series data.In particular, fine-tuning pretrained large language models (LLMs) for time series classification tasks has achieved state-ofthe-art (SOTA) performance on standard benchmarks.However, these LLM-based models have a significant drawback due to the large model size, with the number of trainable parameters in the millions.In this paper, we propose an alternative approach to leveraging the success of language modeling in the time series domain.Instead of fine-tuning LLMs, we utilize a text embedding model to embed time series and then pair the embeddings with a simple classification head composed of convolutional neural networks (CNN) and multilayer perceptron (MLP).We conducted extensive experiments on a well-established time series classification benchmark.We demonstrated LETS-C not only outperforms the current SOTA in classification accuracy but also offers a lightweight solution, using only 14.5% of the trainable parameters on average compared to the SOTA model.Our findings suggest that leveraging text embedding models to encode time series data, combined with a simple yet effective classification head, offers a promising direction for achieving high-performance time series classification while maintaining a lightweight model architecture.

State-of-the-art multimodal web agents, powered by Multimodal Large Language Models (MLLMs), can autonomously execute many web tasks by processing user instructions and interacting with graphical user interfaces (GUIs). Current strategies for building web agents rely on (i) the generalizability of underlying MLLMs and their steerability via prompting, and (ii) large-scale fine-tuning of MLLMs on web-related tasks. However, web agents still struggle to automate tasks on unseen websites and domains, limiting their applicability to enterprise-specific and proprietary platforms. Beyond generalization from large-scale pre-training and fine-tuning, we propose building agents for few-shot adaptability using human demonstrations. We introduce the AdaptAgent framework that enables both proprietary and open-weights multimodal web agents to adapt to new websites and domains using few human demonstrations (up to 2). Our experiments on two popular benchmarks — Mind2Web & VisualWebArena — show that using in-context demonstrations (for proprietary models) or meta-adaptation demonstrations (for meta-learned open-weights models) boosts task success rate by 3.36% to 7.21% over non-adapted state-of-the-art models, corresponding to a relative increase of 21.03% to 65.75%. Furthermore, our additional analyses (a) show the effectiveness of multimodal demonstrations over text-only ones, (b) illuminate how different meta-learning data selection strategies influence the agent’s generalization, and (c) demonstrate how the number of few-shot examples affects the web agent’s success rate. Our results offer a complementary axis for developing widely applicable multimodal web agents beyond large-scale pre-training and fine-tuning, emphasizing few-shot adaptability.

Large Language Models (LLMs) offer the potential for automatic time series analysis and reporting, which is a critical task across many domains, spanning healthcare, finance, climate, energy, and many more.In this paper, we propose a framework for rigorously evaluating the capabilities of LLMs on time series understanding, encompassing both univariate and multivariate forms.We introduce a comprehensive taxonomy of time series features, a critical framework that delineates various characteristics inherent in time series data.Leveraging this taxonomy, we have systematically designed and synthesized a diverse dataset of time series, embodying the different outlined features, each accompanied by textual descriptions.This dataset acts as a solid foundation for assessing the proficiency of LLMs in comprehending time series.Our experiments shed light on the strengths and limitations of stateof-the-art LLMs in time series understanding, revealing which features these models readily comprehend effectively and where they falter.In addition, we uncover the sensitivity of LLMs to factors including the formatting of the data, the position of points queried within a series and the overall time series length.Table 1: Taxonomy of time series characteristics.

This study examined the effectiveness of a vision-based framework for multiple sclerosis (MS) and Parkinson's disease (PD) gait dysfunction prediction. We collected gait video data from multi-view digital cameras during self-paced walking from MS, PD patients and age, weight, height and gender-matched healthy older adults (HOA). We then extracted characteristic 3D joint keypoints from the collected videos. In this work, we proposed a data-driven methodology to classify strides in persons with MS (PwMS), persons with PD (PwPD) and HOA that may generalize across different walking tasks and subjects. We presented a comprehensive quantitative comparison of 16 diverse traditional machine and deep learning (DL) algorithms. When generalizing from comfortable walking (W) to walking-while-talking (WT), multi-scale residual neural network achieved perfect accuracy and AUC for classifying individuals with a given gait disorder; for subject generalization in W trials, residual neural network resulted in the highest accuracy and AUC of 78.1% and 0.87 (resp.), and 1D convolutional neural network (CNN) had highest accuracy of 75% in WT trials. Finally, when generalizing over new subjects in different tasks, again 1D CNN had the top classification accuracy and AUC of 79.3% and 0.93 (resp.). This work is the first attempt to apply and demonstrate the potential of DL with a multi-view digital camera-based gait analysis framework for neurological gait dysfunction prediction. This study suggests the viability of inexpensive vision-based systems for diagnosing certain neurological disorders.

We study the interaction between balance function and anxiety in humans via a virtual reality based experimental setup, fundamentally designed to examine postural control and neurological feedback during balance-demanding height and perturbation shifts, while in quiet stance. In this work, sensory organization balance test (SOT) was utilized to quantify the balance function in subjects, and the salient features of the recorded electrocardiogram (EKG), electromyogram (EMG) and electroencephalography (EEG) signals described the subject's physiological anxiety-related responses. Also, patient-reported outcomes measurement information system (PROMIS) trait anxiety questionnaire was adopted to indicate the self-assessed trait anxiety scores by the subjects. Our statistical analysis indicates that subjects with higher SOT scores encounter a lower heart rate variability during baseline recording relative to the experimental trials and lower co-contraction index of the ankle dorsiflexion (during the experimental trials), suggesting a lower anxiety state. A similar trend was further observed via the self-reported PROMIS questionnaire responses, estimating anxiety experienced in daily life. Additionally, we utilize machine learning to describe the predictability of balance function states via the relative α-band power from the frontal channels of the recorded EEG signals as descriptive features characterizing anxiety. The discussed trends analytically demonstrate the potential for adopting balance function to anticipate anxiety, especially in stressful environments involving postural threats.