Autism Spectrum Disorder (ASD) affects 1 in 54 children in the United States. A core social communication skill negatively impacted by ASD is joint attention (JA), which influences the development of language, cognitive, and social skills from infancy onward. Although several technology-based JA studies have shown potential, they primarily focus on response to joint attention (RJA). The other important component of JA, the initiation of joint attention (IJA), has received less attention from a technology-based intervention perspective. In this work, we present an immersive Computer-mediated Caregiver-Child Interaction (C3I) system to help children with ASD practice IJA skills. C3I is a novel computerized intervention system that integrates a caregiver in the teaching loop, thereby preserving the advantages of both human and computer-administered intervention. A feasibility study with 6 dyads (caregiver-child with ASD) was conducted. A near significant increase with medium effect size on IJA performance was observed. Meanwhile, physiology-based stress analysis showed that C3I did not increase stress of the caregivers over the course of the study. To the best of our knowledge, this is the first autonomous system designed for teaching IJA skills to children with ASD incorporating caregivers within the loop to enhance the potential for generalization in real-world.

Multi-agent collaboration is the latest trending method to build conversational recommender systems (CRS), especially with the widespread use of Large Language Models (LLMs) recently. Typically, these systems employ several LLM agents, each serving distinct roles to meet user needs. In an industrial setting, it’s essential for a CRS to exhibit low first token latency (i.e., the time taken from a user’s input until the system outputs its first response token.) and high scalability—for instance, minimizing the number of LLM inferences per user request—to enhance user experience and boost platform profit. For example, JD.com’s baseline CRS features two LLM agents and a search API but suffers from high first token latency and requires two LLM inferences per request (LIPR), hindering its performance. To address these issues, we introduce a Hybrid Multi-Agent Collaborative Recommender System (Hybrid-MACRS). It includes a central agent powered by a fine-tuned proprietary LLM and a search agent combining a related search module with a search engine. This hybrid system notably reduces first token latency by about 70% and cuts the LIPR from 2 to 1. We conducted thorough online A/B testing to confirm this approach’s efficiency.

Large-scale search and recommendation systems typically adopt a cascaded architecture of retrieval, pre-ranking, ranking, and re-ranking to balance efficiency and accuracy. However, pre-ranking still faces challenges of behavioral sparsity, limited interest diversity, and computational latency. We propose the Query-Attention Dual-Stream (QADS) framework to address these issues. QADS partitions user behaviors into strongly and weakly correlated streams and further decomposes them into fine-grained subsequences guided by domain knowledge. A query-centric attention mechanism reduces complexity from O(N) to O(1), enabling efficient inter- and intra-sequence modeling. A contrastive cross-category transfer module propagates dense patterns from weakly correlated to sparse domains, while a latency-aware parallel inference architecture further reduces delay by 36%. Experiments on public and industrial datasets show that QADS delivers significant performance improvements and has been successfully deployed in large-scale e-commerce search systems.

Generative retrieval formulates document retrieval as an identifier generation task. While prevailing methods increasingly adopt Semantic IDs (SIDs), their opaque nature and rigid mappings struggle with the dynamic inventory and strict interpretability requirements of E-commerce search. Furthermore, generating accurate targets from brief queries against noisy, loosely structured item titles remains a practical challenge. To address these issues, we propose a Query-to-Multi-Span generative retrieval framework tailored for E-commerce. Instead of relying on opaque SIDs or raw titles, our method simplifies the process by generating interpretable multispan identifiers from queries. We align the autoregressive model with user preferences using click logs, and employ a constraintbased beam search to isolate key spans for final item retrieval. This approach explicitly bridges generative models with robust constraint matching, ensuring both matching accuracy and transparency. Extensive offline evaluations demonstrate competitive retrieval performance, and online A/B tests confirm its effectiveness in delivering measurable conversion gains in a production environment.