Raymond Chi-Wing Wong

Generating executable structured queries for OpenStreetMap (OSM) from natural language remains a challenging task due to the rigid syntax of OverpassQL and the steep learning curve it presents to end-users. While steering Large Language Models (LLMs) via demonstration contexts offers a promising solution, existing retrieval-based approaches typically neglect the critical constraint of token budgets. This oversight often leads to prohibitive inference costs and ignores the combinatorial nature of context selection, resulting in suboptimal performance. Furthermore, the lack of reliable benchmarks hinders rigorous evaluation in this domain. To address these challenges, we first introduce OsmNL , a robust benchmark curated via a human-in-the-loop pipeline to resolve alignment noise in prior datasets. Building upon this foundation, we formulate the construction of optimal demonstration contexts under constraints as a Budgeted Example Subset Mining ( BESM ) problem. We analyze BESM as an NP-hard optimization task that admits a submodular objective, theoretically balancing semantic relevance with structural diversity. Leveraging this property, we propose TACO ( T oken- A ware C ontext O ptimization), a principled framework that instantiates a family of budget-aware algorithms to navigate the trade-off between approximation guarantees and computational efficiency. Extensive experiments on OsmNL show that TACO establishes a new state-of-the-art, achieving superior accuracy compared to strong baselines while reducing token consumption by over 50%, thereby offering a robust and cost-efficient solution for large-scale geospatial query interfaces.

Directed graphs are essential for modeling real-world networks, where edge directionality encodes critical information. Cohesive subgraphs in directed graphs, such as directed $(k,\ell )$-plexes (DPlexes), are widely used in applications including community detection, anomaly detection, and functional group discovery. A DPlex is defined as a subgraph in which each vertex disconnects at most $k$ vertices and is meanwhile not pointed to by at most $\ell$ vertices. However, enumerating all maximal DPlexes is computationally challenging, as finding even one maximum DPlex is NP-hard. Furthermore, existing algorithms for undirected graphs cannot be directly applied to directed graphs due to the complexities introduced by edge directionality. To address these challenges, we propose an efficient branch-and-bound algorithm, DPEnum, which incorporates two key techniques: (1) a pivot technique based on a new concept, out-degeneracy ($\delta$), which quantifies sparseness in directed graphs and satisfies $\delta \lt n$, where $n$ is the number of vertices; and (2) a preprocessing technique that removes reducible vertices and directly identifies several maximal DPlexes. DPEnum achieves a worst-case time complexity of $O^*(\alpha _{k}^\delta )$, where $O^*$ suppresses the polynomial factors and $\alpha _{k} \lt 2$, outperforming the naive $O^*(2^{n})$. Additionally, we propose LDPEnum for enumerating large maximal DPlexes with at least $s$ vertices, which uses the upper-bounding technique to prune search branches. Extensive experiments on real-world datasets show that our algorithms achieve up to four orders of magnitude speedup compared to the baseline.

Recent advances in LLM-based agent systems enable natural language–driven, cross-modal data visualization. Yet, existing systems remain unreliable for complex, multi-modal, or iterative tasks. This demonstration introduces MultiVis-Agent, an LLM agent-based system enhanced with logic rules for reliable visualization generation. It integrates specialized agents for data querying, visualization, and validation under a central coordinator governed by a four-layer logic rule framework ensuring parameter safety, error recovery, and guaranteed termination. Built upon the MultiVis task—covering four scenarios from basic generation to iterative refinement—MultiVis-Agent produces executable visualization code from natural language, databases, and optional references. We will show that MultiVis-Agent achieves real-time generation, execution, and refinement with 99% task completion and strong cross-modal reasoning. In particular, we emphasize an end-to-end and interactive demonstration where users can upload databases, provide multi-modal references (images/code), and iteratively refine executable visualization code within the same interface. The demo video and code are available at https://github.com/Jinwei-Lu/MultiVis.

Causality has been integrated with machine learning in uncovering and understanding the causal relationship between variables and observed outcomes. However, the centralized training setting of causal machine learning is not adaptable to most practical scenarios, where datasets are distributed, stored, and unsharable due to privacy concerns. Federated learning (FL), a distributed learning framework that allows collaborative training across multiple devices without raw data sharing, emerges as a potential solution to this problem. By integrating FL into causal problems, the discovery and inference of causal relationships across dispersed datasets can be achieved. On the other hand, causality can also enhance FL models in various dimensions, including model interpretability and explainability, generalizability, adversarial robustness, and fairness and bias mitigation. In this article, we provide a comprehensive review of the above two directions and summarize the interplays between causality and FL (short for Causal-FL ) by organizing our discussion around two key questions: (1) how FL enable decentralized causal analysis; and (2) how causality tackles FL challenges. The potential applications of these methods are also introduced, including healthcare, recommendation, economics, social equity, and so on. Moreover, we discuss promising future directions and future challenges to be explored.

Real-world visualization tasks involve complex, multi-modal requirements that extend beyond simple text-to-chart generation, requiring reference images, code examples, and iterative refinement. Current systems exhibit fundamental limitations: single-modality input, one-shot generation, and rigid workflows. While LLM-based approaches show potential for these complex requirements, they introduce reliability challenges including catastrophic failures and infinite loop susceptibility. To address this gap, we propose MultiVis-Agent, a logic rule-enhanced multi-agent framework for reliable multi-modal and multi-scenario visualization generation. Our approach introduces a four-layer logic rule framework that provides mathematical guarantees for system reliability while maintaining flexibility. Unlike traditional rule-based systems, our logic rules are mathematical constraints that guide LLM reasoning rather than replacing it. We formalize the MultiVis task spanning four scenarios from basic generation to iterative refinement, and develop MultiVis-Bench, a benchmark with over 1,000 cases for multi-modal visualization evaluation. Extensive experiments demonstrate that our approach achieves 75.63% visualization score on challenging tasks, significantly outperforming baselines (57.54-62.79%), with task completion rates of 99.58% and code execution success rates of 94.56% (vs. 74.48% and 65.10% without logic rules), successfully addressing both complexity and reliability challenges in automated visualization generation.

Time series data are ubiquitous across many different fields. Many data mining tasks, such as classification, clustering, and motif finding, have been defined for time series data. They use similarity measures as core subroutines, making it crucial to design them to align with our intuitions. Dynamic Time Warping (DTW) is arguably the most prevailing distance measure. However, studies have shown that for certain data, another distance measure, namely Uniform Scaling (US), is equally crucial as DTW. DTW handles the local distortion, while US handles the global scaling. In addition, studies have demonstrated that combining DTW and US is necessary in some cases. Surprisingly, all existing studies assume a single scaling factor across the entire time series. A time series could consist of phases. Since each phase expresses at its own rate, using a single scaling factor is insufficient when comparing two time series that share similar phases but differ in expression rate. We introduce the first framework that accounts for multiple scaling factors, Piecewise Scaling Distance (PSD). It employs other distance measures as its base distance measure. Because the direct implementation is slow, we propose a constrained version that enforces constraints based on the allowed segment lengths derived from the given scaling factor bound. It also prevents pathological results. Two additional speedup techniques have been proposed, which achieve 8.59X to 176.14X speedup. We also demonstrate the use of a lower bound when DTW is employed as the base distance measure to accelerate PSD computation. Moreover, we show that PSD segmentation results can improve the accuracy of other distance measures.

Time Series Foundation Models (TSFMs) achieve strong zero-shot forecasting through large-scale pre-training, but adapting them to downstream domains under distribution shift remains challenging. Existing solutions face a trade-off: Parametric Adaptation can cause catastrophic forgetting and requires costly multi-domain maintenance, while Non-Parametric Retrieval improves forecasts but incurs high inference latency due to datastore search. We propose Parametric Memory Distillation and implement it as TS-Memory, a lightweight memory adapter that augments frozen TSFMs. TS-Memory is trained in two stages. First, we construct an offline, leakage-safe kNN teacher that synthesizes confidence-aware quantile targets from retrieved futures. Second, we distill this retrieval-induced distributional correction into a lightweight memory adapter via confidence-gated supervision. During inference, TS-Memory fuses memory and backbone predictions with constant-time overhead, enabling retrieval-free deployment. Experiments across diverse TSFMs and benchmarks demonstrate consistent improvements in both point and probabilistic forecasting over representative adaptation methods, with efficiency comparable to the frozen backbone.