Modern recommender systems operate in uniquely dynamic settings: user interests, item pools, and popularity trends shift continuously, and models must adapt in real time without forgetting past preferences. While existing tutorials on continual or lifelong learning cover broad machine learning domains (e.g., vision and graphs), they do not address recommendation-specific demands-such as balancing stability and plasticity per user, handling cold-start items, and optimizing recommendation metrics under streaming feedback. This tutorial aims to make a timely contribution by filling that gap. We begin by reviewing the background and problem settings, followed by a comprehensive overview of existing approaches. We then highlight recent efforts to apply continual learning to practical deployment environments, such as resource-constrained systems and sequential interaction settings. Finally, we discuss open challenges and future research directions. We expect this tutorial to benefit researchers and practitioners in recommender systems, data mining, AI, and information retrieval across academia and industry.

Our approach is motivated by the potential of leveraging noisy synthetic data from driving world models to enhance BEV semantic segmentation. The proposed method investigates a noise-resilient learning framework designed for handling synthetic data with inherent noise. The generated data from different world models exhibits inconsistent road structures at identical viewpoints. Abstract --Birds' Eye View (BEV) semantic segmentation is an indispensable perception task in end-to-end autonomous driving systems. Unsupervised and semi-supervised learning for BEV tasks, as pivotal for real-world applications, underperform due to the homogeneous distribution of the labeled data. In this work, we explore the potential of synthetic data from driving world models to enhance the diversity of labeled data for robustifying BEV segmentation. Y et, our preliminary findings reveal that generation noise in synthetic data compromises efficient BEV model learning. T o fully harness the potential of synthetic data from world models, this paper proposes NRSeg, a noise-resilient learning framework for BEV semantic segmentation. Specifically, a Perspective-Geometry Consistency Metric (PGCM) is proposed to quantitatively evaluate the guidance capability of generated data for model learning. This metric originates from the alignment measure between the perspective road mask of generated data and the mask projected from the BEV labels. This work was supported in part by the National Natural Science Foundation of China (No. U21A20518, No. 61976086, and No. 62473139) and in part by the Open Research Project of the State Key Laboratory of Industrial Control Technology, China (Grant No. ICT2025B20). Wang are with the School of Robotics and the National Engineering Research Center of Robot Visual Perception and Control Technology, Hunan University, Changsha 410082, China (email: kailun.yang@hnu.edu.cn; Cao is with the Key Laboratory of Big Data Research and Application for Basic Education, Hunan Normal University, Changsha 410006, China.

Traditional indoor robot navigation methods provide a reliable solution when adapted to constrained scenarios, but lack flexibility or require manual re-tuning when deployed in more complex settings. In contrast, learning-based approaches learn directly from sensor data and environmental interactions, enabling easier adaptability. While significant work has been presented in the context of learning navigation policies, learning-based methods are rarely compared to traditional navigation methods directly, which is a problem for their ultimate acceptance in general navigation contexts. In this work, we explore the viability of imitation learning (IL) for indoor navigation, using expert (joystick) demonstrations to train various navigation policy networks based on RGB images, LiDAR, and a combination of both, and we compare our IL approach to a traditional potential field-based navigation method. We evaluate the approach on a physical mobile robot platform equipped with a 2D LiDAR and a camera in an indoor university environment. Our multimodal model demonstrates superior navigation capabilities in most scenarios, but faces challenges in dynamic environments, likely due to limited diversity in the demonstrations. Nevertheless, the ability to learn directly from data and generalise across layouts suggests that IL can be a practical navigation approach, and potentially a useful initialisation strategy for subsequent lifelong learning.

With the rapid progress of artificial intelligence (AI) in multi-modal understanding, there is increasing potential for video comprehension technologies to support professional domains such as medical education. However, existing benchmarks suffer from two primary limitations: (1) Linguistic Singularity: they are largely confined to English, neglecting the need for multilingual resources; and (2) Shallow Reasoning: their questions are often designed for surface-level information retrieval, failing to properly assess deep multi-modal integration. To address these limitations, we present M3-Med, the first benchmark for Multi-lingual, Multi-modal, and Multi-hop reasoning in Medical instructional video understanding. M3-Med consists of medical questions paired with corresponding video segments, annotated by a team of medical experts. A key innovation of M3-Med is its multi-hop reasoning task, which requires a model to first locate a key entity in the text, then find corresponding visual evidence in the video, and finally synthesize information across both modalities to derive the answer. This design moves beyond simple text matching and poses a substantial challenge to a model's deep cross-modal understanding capabilities. We define two tasks: Temporal Answer Grounding in Single Video (TAGSV) and Temporal Answer Grounding in Video Corpus (TAGVC). We evaluated several state-of-the-art models and Large Language Models (LLMs) on M3-Med. The results reveal a significant performance gap between all models and human experts, especially on the complex multi-hop questions where model performance drops sharply. M3-Med effectively highlights the current limitations of AI models in deep cross-modal reasoning within specialized domains and provides a new direction for future research.

Religion and spirituality (R/S) are complex and highly domain-dependent concepts which have long confounded researchers and policymakers. Due to their context-specificity, R/S are difficult to operationalize in conventional archival search strategies, particularly when datasets are very large, poorly accessible, and marked by information noise. As a result, considerable time investments and specialist knowledge is often needed to extract actionable insights related to R/S from general archival sources, increasing reliance on published literature and manual desk reviews. To address this challenge, we present SpiritRAG, an interactive Question Answering (Q&A) system based on Retrieval-Augmented Generation (RAG). Built using 7,500 United Nations (UN) resolution documents related to R/S in the domains of health and education, SpiritRAG allows researchers and policymakers to conduct complex, context-sensitive database searches of very large datasets using an easily accessible, chat-based web interface. SpiritRAG is lightweight to deploy and leverages both UN documents and user provided documents as source material. A pilot test and evaluation with domain experts on 100 manually composed questions demonstrates the practical value and usefulness of SpiritRAG.

A long-standing problem in online reinforcement learning (RL) is of ensuring sample efficiency, which stems from an inability to explore environments efficiently. Most attempts at efficient exploration tackle this problem in a setting where learning begins from scratch, without prior information available to bootstrap learning. However, such approaches fail to leverage expert demonstrations and simulators that can reset to arbitrary states. These affordances are valuable resources that offer enormous potential to guide exploration and speed up learning. In this paper, we explore how a small number of expert demonstrations and a simulator allowing arbitrary resets can accelerate learning during online RL. We find that training with a suitable choice of an auxiliary start state distribution that may differ from the true start state distribution of the underlying Markov Decision Process can significantly improve sample efficiency. We find that using a notion of safety to inform the choice of this auxiliary distribution significantly accelerates learning. By using episode length information as a way to operationalize this notion, we demonstrate state-of-the-art sample efficiency on a sparse-reward hard-exploration environment.

In the era of mobile computing, deploying efficient Natural Language Processing (NLP) models in resource-restricted edge settings presents significant challenges, particularly in environments requiring strict privacy compliance, real-time responsiveness, and diverse multi-tasking capabilities. These challenges create a fundamental need for ultra-compact models that maintain strong performance across various NLP tasks while adhering to stringent memory constraints. To this end, we introduce Edge ultra-lIte BERT framework (EI-BERT) with a novel cross-distillation method. EI-BERT efficiently compresses models through a comprehensive pipeline including hard token pruning, cross-distillation and parameter quantization. Specifically, the cross-distillation method uniquely positions the teacher model to understand the student model's perspective, ensuring efficient knowledge transfer through parameter integration and the mutual interplay between models. Through extensive experiments, we achieve a remarkably compact BERT-based model of only 1.91 MB - the smallest to date for Natural Language Understanding (NLU) tasks. This ultra-compact model has been successfully deployed across multiple scenarios within the Alipay ecosystem, demonstrating significant improvements in real-world applications. For example, it has been integrated into Alipay's live Edge Recommendation system since January 2024, currently serving the app's recommendation traffic across \textbf{8.4 million daily active devices}.

Large Language Models (LLMs) often generate responses that are factually incorrect yet expressed with high confidence, which can pose serious risks for end users. To address this, it is essential for LLMs not only to produce answers but also to provide accurate estimates of their correctness. Uncertainty quantification methods have been introduced to assess the quality of LLM outputs, with factual accuracy being a key aspect of that quality. Among these methods, those that leverage hidden states to train probes have shown particular promise, as these internal representations encode information relevant to the factuality of responses, making this approach the focus of this paper. However, the probe trained on the hidden states of one dataset often struggles to generalise to another dataset of a different task or domain. To address this limitation, we explore combining data-agnostic features with hidden-state features and assess whether this hybrid feature set enhances out-of-domain performance. We further examine whether selecting only the most informative hidden-state features, thereby discarding task-specific noise, enables the data-agnostic features to contribute more effectively. The experiment results indicate that although introducing data-agnostic features generally enhances generalisation performance in most cases, in certain scenarios their inclusion degrades performance. A similar pattern emerges when retaining only the most important hidden-state features - adding data-agnostic features does not consistently further enhance performance compared to using the full set of hidden-state features. A closer analysis reveals that, in some specific cases, the trained probe underweights the data-agnostic features relative to the hidden-state features, which we believe is the main reason why the results are inconclusive.

There has been a growing interest in enhancing the mathematical problem-solving (MPS) capabilities of large language models. While the majority of research efforts concentrate on creating specialized models to solve mathematical problems, it remains unknown how learning mathematical problem-solving generalizes to help develop other reasoning abilities. In this paper, we present an empirical investigation into the generalization potential of various MPS training approaches, such as continual pretraining, instruction tuning, and rule-based reinforcement learning across various data sources, including both short and long chain-of-thought (CoT) samples. Evaluation on 5 mathematical and 8 general reasoning benchmarks show that continual pretraining on math text is able to generalize to general reasoning tasks to some extent. In constrast, instruction tuning on conventional, short MPS samples provides limited benefits and, in many cases, even impairs generalization performance. Notably, training with long CoT responses for MPS samples and incorporating rule-based reinforcement learning on MPS queries exhibit distinct behavior, significantly enhancing generalization by extending the model's reasoning processes into other domains. These results suggest that traditional approaches to learning MPS with short reasoning chains largely fail to achieve robust generalization. However, the emerging paradigm of longer reasoning chains, coupled with self-reflection, offers a promising direction for improving generalized reasoning abilities through learning from specialized domains.

While chain-of-thought (CoT) monitoring is an appealing AI safety defense, recent work on "unfaithfulness" has cast doubt on its reliability. These findings highlight an important failure mode, particularly when CoT acts as a post-hoc rationalization in applications like auditing for bias. However, for the distinct problem of runtime monitoring to prevent severe harm, we argue the key property is not faithfulness but monitorability. To this end, we introduce a conceptual framework distinguishing CoT-as-rationalization from CoT-as-computation. We expect that certain classes of severe harm will require complex, multi-step reasoning that necessitates CoT-as-computation. Replicating the experimental setups of prior work, we increase the difficulty of the bad behavior to enforce this necessity condition; this forces the model to expose its reasoning, making it monitorable. We then present methodology guidelines to stress-test CoT monitoring against deliberate evasion. Applying these guidelines, we find that models can learn to obscure their intentions, but only when given significant help, such as detailed human-written strategies or iterative optimization against the monitor. We conclude that, while not infallible, CoT monitoring offers a substantial layer of defense that requires active protection and continued stress-testing.