沈骋骜 ChengAo Shen

Time series AI is crucial for analyzing dynamic web content, driving a surge of pre-trained large models known for their strong knowledge encoding and transfer capabilities across diverse tasks. However, given their energy-intensive training, inference, and hardware demands, using large models as a one-fits-all solution raises serious concerns about carbon footprint and sustainability. For a specific task, a compact yet specialized, high-performing model may be more practical and affordable, especially for resource-constrained users such as small businesses. This motivates the question: Can we build cost-effective lightweight models with large-model-like performance on core tasks such as forecasting? This paper addresses this question by introducing SVTime, a novel Small model inspired by large Vision model (LVM) forecasters for long-term Time series forecasting (LTSF). Recently, LVMs have been shown as powerful tools for LTSF. We identify a set of key inductive biases of LVM forecasters – analogous to the “physics” governing their behaviors in LTSF – and design small models that encode these biases through meticulously crafted linear layers and constraint functions. Across 21 baselines spanning lightweight, complex, and pre-trained large models on 8 benchmark datasets, SVTime outperforms state-of-the-art (SOTA) lightweight models and rivals large models with 10^3 fewer parameters than LVMs, while enabling efficient training and inference in low-resource settings.

Time series forecasting (TSF) is evolving toward multimodal and agentic settings, yet using foundation models remains uneconomical in resource-constrained scenarios, where compact, specialized forecasters are more desirable. However, lightweight forecasters typically require substantial training data, limiting their use in domains with scarce, slowly accumulated, or privacy-sensitive time series. To address this dilemma, we investigate the challenging problem of few-shot learning for lightweight forecasters. We propose MetaCaster, a meta-harness-optimized multi-agent framework that uses agentic data generation to automatically train specialized lightweight forecasters from only a few examples and textual contexts. Our work highlights a new TSF paradigm in which agents act not as forecasters but as intermediary engineers that prepare efficient, task-specific forecasters for deployment. Experiments on 18 datasets, 23 state-of-the-art lightweight forecasters, and 14 baselines demonstrate that MetaCaster achieves both data efficiency and computational efficiency while maintaining high-quality TSF performance.

Large Vision Models (LVMs) are emerging tools for transferring cross-modal knowledge to time series, but their potential for this domain is not yet fully understood. This work addresses the gap by investigating LVMs for both high-level (classification) and low-level (forecasting) time series tasks. Our aim is not only to assess whether LVMs can succeed, but also to reveal why they succeed or fall short. Through a comparative benchmark covering four representative LVMs, eight imaging methods, 18 datasets, and 21 baselines, we identify the strengths and limitations of the foundational LVMs, as well as effective strategies for adapting them to time series modeling. Our findings indicate that while the LVMs are effective for time series classification, they face notable challenges in forecasting. In particular, the best-performing LVM-based forecaster is limited to specific model types and imaging methods, exhibits biases toward periodic components in time series, and struggles to leverage long look-back windows. We hope our findings will serve as both a cornerstone and a practical guide for advancing future research on LVM- and multimodal-based solutions for different time series tasks.

Time series, typically represented as numerical sequences, can also be transformed into images and texts, offering multi-modal views (MMVs) of the same underlying signal. These MMVs can reveal complementary patterns and enable the use of powerful pre-trained large models, such as large vision models (LVMs), for long-term time series forecasting (LTSF). However, as we identified in this work, applying LVMs to LTSF poses an inductive bias towards “forecasting periods”. To harness this bias, we propose DMMV, a novel decomposition-based multi-modal view framework that leverages trend-seasonal decomposition and a novel backcast residual based adaptive decomposition to integrate MMVs for LTSF. Comparative evaluations against 14 state-of-the-art (SOTA) models across diverse datasets show that DMMV outperforms single-view and existing multi-modal baselines, achieving the best mean squared error (MSE) on 6 out of 8 benchmark datasets.

Causal discovery is an imperative foundation for decision-making across domains, such as smart health, AI for drug discovery and AIOps. Traditional statistical causal discovery methods, while well-established, predominantly rely on observational data and often overlook the semantic cues inherent in cause-and-effect relationships. The advent of Large Language Models (LLMs) has ushered in an affordable way of leveraging the semantic cues for knowledge-driven causal discovery, but the development of LLMs for causal discovery lags behind other areas, particularly in the exploration of multi-modal data. To bridge the gap, we introduce MATMCD, a multi-agent system powered by tool-augmented LLMs. MATMCD has two key agents: a Data Augmentation agent that retrieves and processes modality-augmented data, and a Causal Constraint agent that integrates multi-modal data for knowledge-driven reasoning. The proposed design of the inner-workings ensures successful cooperation of the agents. Our empirical study across seven datasets suggests the significant potential of multi-modality enhanced causal discovery.

Time series analysis has witnessed the inspiring development from traditional autoregressive models, deep learning models, to recent Transformers and Large Language Models (LLMs). Efforts in leveraging vision models for time series analysis have also been made along the way but are less visible to the community due to the predominant research on sequence modeling in this domain. However, the discrepancy between continuous time series and the discrete token space of LLMs, and the challenges in explicitly modeling the correlations of variates in multivariate time series have shifted some research attentions to the equally successful Large Vision Models (LVMs) and Vision Language Models (VLMs). To fill the blank in the existing literature, this survey discusses the advantages of vision models over LLMs in time series analysis. It provides a comprehensive and in-depth overview of the existing methods, with dual views of detailed taxonomy that answer the key research questions including how to encode time series as images and how to model the imaged time series for various tasks. Additionally, we address the challenges in the pre- and post-processing steps involved in this framework and outline future directions to further advance time series analysis with vision models.

The image fusion method is widely used in many different fields. The fusion processes both need models to extract semantic information and contain details. Traditional image processing techniques used for this issue have limited ability to extract semantic features from images, and advanced deep learning techniques often lose the details. In this work, we propose the Controlled Fusion Network (CFN) that adopts a multi-step progressive generation method and injects control elements at every step. We test the model in the emoji fusion task which accepts various emojis and combines them. We find that the generated emojis sufficiently retain and reasonably combine the semantic information of the input images, while the result images also conform to human intuitive perception. Our source code is released at: https://github.com/ChengAoShen/Emoji_fusion.