Accurate and efficient analysis of multivariate spatio-temporal data is critical in climatology, geology, and sociology applications. Existing models usually assume simple inter-dependence among variables, space, and time, and are computationally expensive. We propose a unified low rank tensor learning framework for multivariate spatio-temporal analysis, which can conveniently incorporate different properties in spatio-temporal data, such as spatial clustering and shared structure among variables. We demonstrate how the general framework can be applied to cokriging and forecasting tasks, and develop an efficient greedy algorithm to solve the resulting optimization problem with convergence guarantee. We conduct experiments on both synthetic datasets and real application datasets to demonstrate that our method is not only significantly faster than existing methods but also achieves lower estimation error.

Understanding risk in autonomous driving requires not only perception and prediction, but also high-level reasoning about agent behavior and context. Current Vision Language Models (VLMs)-based methods primarily ground agents in static images and provide qualitative judgments, lacking the spatio-temporal reasoning needed to capture how risks evolve over time. To address this gap, we propose NuRisk, a comprehensive Visual Question Answering (VQA) dataset comprising 2,900 scenarios and 1.1 million agent-level samples, built on real-world data from nuScenes and Waymo, supplemented with safety-critical scenarios from the CommonRoad simulator. The dataset provides Bird-Eye-View (BEV) based sequential images with quantitative, agent-level risk annotations, enabling spatio-temporal reasoning. We benchmark well-known VLMs across different prompting techniques and find that they fail to perform explicit spatio-temporal reasoning, resulting in a peak accuracy of 33% at high latency. To address these shortcomings, our fine-tuned 7B VLM agent improves accuracy to 41% and reduces latency by 75%, demonstrating explicit spatio-temporal reasoning capabilities that proprietary models lacked. While this represents a significant step forward, the modest accuracy underscores the profound challenge of the task, establishing NuRisk as a critical benchmark for advancing spatio-temporal reasoning in autonomous driving.

In classical AI, perception relies on learning state-based representations, while planning, which can be thought of as temporal reasoning over action sequences, is typically achieved through search. We study whether such reasoning can instead emerge from representations that capture both perceptual and temporal structure. We show that standard temporal contrastive learning, despite its popularity, often fails to capture temporal structure due to its reliance on spurious features. To address this, we introduce Combinatorial Representations for Temporal Reasoning (CRTR), a method that uses a negative sampling scheme to provably remove these spurious features and facilitate temporal reasoning. CRTR achieves strong results on domains with complex temporal structure, such as Sokoban and Rubik's Cube. In particular, for the Rubik's Cube, CRTR learns representations that generalize across all initial states and allow it to solve the puzzle using fewer search steps than BestFS, though with longer solutions. To our knowledge, this is the first method that efficiently solves arbitrary Cube states using only learned representations, without relying on an external search algorithm.

Modern BPE tokenizers often split calendar dates into meaningless fragments, e.g., 20250312 $\rightarrow$ 202, 503, 12, inflating token counts and obscuring the inherent structure needed for robust temporal reasoning. In this work, we (1) introduce a simple yet interpretable metric, termed date fragmentation ratio, that measures how faithfully a tokenizer preserves multi-digit date components; (2) release DateAugBench, a suite of 6500 examples spanning three temporal reasoning tasks: context-based date resolution, format-invariance puzzles, and date arithmetic across historical, contemporary, and future time periods; and (3) through layer-wise probing and causal attention-hop analyses, uncover an emergent date-abstraction mechanism whereby large language models stitch together the fragments of month, day, and year components for temporal reasoning. Our experiments show that excessive fragmentation correlates with accuracy drops of up to 10 points on uncommon dates like historical and futuristic dates. Further, we find that the larger the model, the faster the emergent date abstraction that heals date fragments is accomplished. Lastly, we observe a reasoning path that LLMs follow to assemble date fragments, typically differing from human interpretation (year $\rightarrow$ month $\rightarrow$ day). Our datasets and code are made publicly available \href{https://github.com/gagan3012/date-fragments}{here}.

Modeling evolving interactions among entities is critical in many real-world tasks. For example, predicting driver maneuvers in traffic requires tracking how neighboring vehicles accelerate, brake, and change lanes relative to one another over consecutive frames. Likewise, detecting financial fraud hinges on following the flow of funds through successive transactions as they propagate through the network. Unlike classic time-series forecasting, these settings demand reasoning over who interacts with whom and when, calling for a temporal-graph representation that makes both the relations and their evolution explicit. Existing temporal-graph methods typically use snapshot graphs to encode temporal evolution. We introduce a full-history graph that instantiates one node for every entity at every time step and separates two edge sets: (i) intra-time-step edges that capture relations within a single frame and (ii) inter-time-step edges that connect an entity to itself at consecutive steps. To learn on this graph we design an Edge-Type Decoupled Network (ETDNet) with parallel modules: a graph-attention module aggregates information along intra-time-step edges, a multi-head temporal-attention module attends over an entity's inter-time-step history, and a fusion module combines the two messages after every layer. Evaluated on driver-intention prediction (Waymo) and Bitcoin fraud detection (Elliptic++), ETDNet consistently surpasses strong baselines, lifting Waymo joint accuracy to 75.6\% (vs. 74.1\%) and raising Elliptic++ illicit-class F1 to 88.1\% (vs. 60.4\%). These gains demonstrate the benefit of representing structural and temporal relations as distinct edges in a single graph.

We address the task of temporal knowledge graph (TKG) forecasting by introducing a fully explainable method based on temporal rules. Motivated by recent work proposing a strong baseline using recurrent facts, our approach learns four simple types of rules with a confidence function that considers both recency and frequency. Evaluated on nine datasets, our method matches or surpasses the performance of eight state-of-the-art models and two baselines, while providing fully interpretable predictions.

Surgical video understanding is crucial for facilitating Computer-Assisted Surgery (CAS) systems. Despite significant progress in existing studies, two major limitations persist, including inadequate visual content perception and insufficient temporal awareness in surgical videos, and hinder the development of versatile CAS solutions. In this work, we propose the SurgLLM framework, an effective large multimodal model tailored for versatile surgical video understanding tasks with enhanced spatial focus and temporal awareness. Specifically, to empower the spatial focus of surgical videos, we first devise Surgical Context-aware Multimodal Pretraining (Surg-Pretrain) for the video encoder of SurgLLM, by performing instrument-centric Masked Video Reconstruction (MV-Recon) and subsequent multimodal alignment. To incorporate surgical temporal knowledge into SurgLLM, we further propose Temporal-aware Multimodal Tuning (TM-Tuning) to enhance temporal reasoning with interleaved multimodal embeddings. Moreover, to accommodate various understanding tasks of surgical videos without conflicts, we devise a Surgical Task Dynamic Ensemble to efficiently triage a query with optimal learnable parameters in our SurgLLM. Extensive experiments performed on diverse surgical video understanding tasks, including captioning, general VQA, and temporal VQA, demonstrate significant improvements over the state-of-the-art approaches, validating the effectiveness of our SurgLLM in versatile surgical video understanding. The source code is available at https://github.com/franciszchen/SurgLLM.

The supplementary material is structured as follows: Section A.1 gives the proof and analysis of Theorem 3.1; Section A.2 introduces the datasets and their statistics in detail; Section A.3 introduces the baselines utilized in experiments; Section A.4 discusses the experimental setup of baseline models as well as MetaTKGR; Section A.5 reports detailed experiment performance with statistical test results; A.1 Statements, Proof and Analysis of Theorem 3.1 Thus, we can improve the generalization ability of our meta-learner over time by the following update step by step, A.2 Datasets Figure 1: Number of entities over time. New entities continuously emerge on three public TKGs. Integrated Crisis Early Warning System (ICEWS18) is the collection of coded interactions between 3 socio-political actors which are extracted from news articles. Y AGO). Figure 1 shows the amount of new entities appearing over time. Figure 2 shows the corresponding distributions.

Augmented Reality (AR) systems are increasingly integrating foundation models, such as Multimodal Large Language Models (MLLMs), to provide more context-aware and adaptive user experiences. This integration has led to the development of AR agents to support intelligent, goal-directed interactions in real-world environments. While current AR agents effectively support immediate tasks, they struggle with complex multi-step scenarios that require understanding and leveraging user's long-term experiences and preferences. This limitation stems from their inability to capture, retain, and reason over historical user interactions in spatiotemporal contexts. To address these challenges, we propose a conceptual framework for memory-augmented AR agents that can provide personalized task assistance by learning from and adapting to user-specific experiences over time. Our framework consists of four interconnected modules: (1) Perception Module for multimodal sensor processing, (2) Memory Module for persistent spatiotemporal experience storage, (3) Spatiotemporal Reasoning Module for synthesizing past and present contexts, and (4) Actuator Module for effective AR communication. We further present an implementation roadmap, a future evaluation strategy, a potential target application and use cases to demonstrate the practical applicability of our framework across diverse domains. We aim for this work to motivate future research toward developing more intelligent AR systems that can effectively bridge user's interaction history with adaptive, context-aware task assistance.

Graph Retrieval-Augmented Generation has emerged as a powerful paradigm for grounding large language models with external structured knowledge. However, existing Graph RAG methods struggle with temporal reasoning, due to their inability to model the evolving structure and order of real-world events. In this work, we introduce DyG-RAG, a novel event-centric dynamic graph retrieval-augmented generation framework designed to capture and reason over temporal knowledge embedded in unstructured text. To eliminate temporal ambiguity in traditional retrieval units, DyG-RAG proposes Dynamic Event Units (DEUs) that explicitly encode both semantic content and precise temporal anchors, enabling accurate and interpretable time-aware retrieval. To capture temporal and causal dependencies across events, DyG-RAG constructs an event graph by linking DEUs that share entities and occur close in time, supporting efficient and meaningful multi-hop reasoning. To ensure temporally consistent generation, DyG-RAG introduces an event timeline retrieval pipeline that retrieves event sequences via time-aware traversal, and proposes a Time Chain-of-Thought strategy for temporally grounded answer generation. This unified pipeline enables DyG-RAG to retrieve coherent, temporally ordered event sequences and to answer complex, time-sensitive queries that standard RAG systems cannot resolve. Extensive experiments on temporal QA benchmarks demonstrate that DyG-RAG significantly improves the accuracy and recall of three typical types of temporal reasoning questions, paving the way for more faithful and temporal-aware generation. DyG-RAG is available at https://github .com/RingBDStack/DyG-RAG.