Jun Wang

Yuyuan Li, Chaochao Chen, Xiaolin Zheng et al.

With the growing privacy concerns in recommender systems, recommendation unlearning, i.e., forgetting the impact of specific learned targets, is getting increasing attention. Existing studies predominantly use training data, i.e., model inputs, as the unlearning target. However, we find that attackers can extract private information, i.e., gender, race, and age, from a trained model even if it has not been explicitly encountered during training. We name this unseen information as attribute and treat it as the unlearning target. To protect the sensitive attribute of users, Attribute Unlearning (AU) aims to degrade attacking performance and make target attributes indistinguishable. In this paper, we focus on a strict but practical setting of AU, namely Post-Training Attribute Unlearning (PoT-AU), where unlearning can only be performed after the training of the recommendation model is completed. To address the PoT-AU problem in recommender systems, we design a two-component loss function that consists of i) distinguishability loss: making attribute labels indistinguishable from attackers, and ii) regularization loss: preventing drastic changes in the model that result in a negative impact on recommendation performance. Specifically, we investigate two types of distinguishability measurements, i.e., user-to-user and distribution-to-distribution. We use the stochastic gradient descent algorithm to optimize our proposed loss. Extensive experiments on three real-world datasets demonstrate the effectiveness of our proposed methods.

Xinting Liao, Weiming Liu, Chaochao Chen et al.

Federated learning (FL) collaboratively models user data in a decentralized way. However, in the real world, non-identical and independent data distributions (non-IID) among clients hinder the performance of FL due to three issues, i.e., (1) the class statistics shifting, (2) the insufficient hierarchical information utilization, and (3) the inconsistency in aggregating clients. To address the above issues, we propose HyperFed which contains three main modules, i.e., hyperbolic prototype Tammes initialization (HPTI), hyperbolic prototype learning (HPL), and consistent aggregation (CA). Firstly, HPTI in the server constructs uniformly distributed and fixed class prototypes, and shares them with clients to match class statistics, further guiding consistent feature representation for local clients. Secondly, HPL in each client captures the hierarchical information in local data with the supervision of shared class prototypes in the hyperbolic model space. Additionally, CA in the server mitigates the impact of the inconsistent deviations from clients to server. Extensive studies of four datasets prove that HyperFed is effective in enhancing the performance of FL under the non-IID set.

Filippos Christianos, Georgios Papoudakis, Thomas Coste et al.

This paper introduces a novel mobile phone control architecture, Lightweight Multi-modal App Control (LiMAC), for efficient interactions and control across various Android apps. LiMAC takes as input a textual goal and a sequence of past mobile observations, such as screenshots and corresponding UI trees, to generate precise actions. To address the computational constraints inherent to smartphones, we introduce a small Action Transformer (AcT) integrated with a fine-tuned vision-language model (VLM) for real-time decision-making and task execution. We evaluate LiMAC on two open-source mobile control datasets, demonstrating the superior performance of our small-form-factor approach against fine-tuned versions of open-source VLMs, such as Florence2 and Qwen2-VL. It also significantly outperforms prompt engineering baselines utilising closed-source foundation models like GPT-4o. More specifically, LiMAC increases the overall action accuracy by up to 19% compared to fine-tuned VLMs, and up to 42% compared to prompt-engineering baselines.

Heyuan Huang, Xingyu Lou, Chaochao Chen et al.

Cross-Domain Recommendation (CDR) have received widespread attention due to their ability to utilize rich information across domains. However, most existing CDR methods assume an ideal static condition that is not practical in industrial recommendation systems (RS). Therefore, simply applying existing CDR methods in the industrial RS environment may lead to low effectiveness and efficiency. To fill this gap, we propose DIIT, an end-to-end Domain-Invariant Information Transfer method for industrial cross-domain recommendation. Specifically, We first simulate the industrial RS environment that maintains respective models in multiple domains, each of them is trained in the incremental mode. Then, for improving the effectiveness, we design two extractors to fully extract domain-invariant information from the latest source domain models at the domain level and the representation level respectively. Finally, for improving the efficiency, we design a migrator to transfer the extracted information to the latest target domain model, which only need the target domain model for inference. Experiments conducted on one production dataset and two public datasets verify the effectiveness and efficiency of DIIT.

Heng Tang, Feng Liu, Xinbo Chen et al.

Recent years have witnessed extensive exploration of Large Language Models (LLMs) on the field of Recommender Systems (RS). There are currently two commonly used strategies to enable LLMs to have recommendation capabilities: 1) The "Guidance-Only" strategy uses in-context learning to exploit and amplify the inherent semantic understanding and item recommendation capabilities of LLMs; 2) The "Tuning-Only" strategy uses supervised fine-tuning (SFT) to fine-tune LLMs with the aim of fitting them to real recommendation data. However, neither of these strategies can effectively bridge the gap between the knowledge space of LLMs and recommendation, and their performance do not meet our expectations. To better enable LLMs to learn recommendation knowledge, we combine the advantages of the above two strategies and proposed a novel "Guidance+Tuning" method called Self-Optimized Fine-Tuning (SOFT), which adopts the idea of curriculum learning. It first employs self-distillation to construct an auxiliary easy-to-learn but meaningful dataset from a fine-tuned LLM. Then it further utilizes a self-adaptive curriculum scheduler to enable LLMs to gradually learn from simpler data (self-distilled data) to more challenging data (real RS data). Extensive experiments demonstrate that SOFT significantly enhances the recommendation accuracy (37.59\% on average) of LLM-based methods. The code is available via https://anonymous.4open.science/r/Self-Optimized-Fine-Tuning-264E

Adam X. Yang, Maxime Robeyns, Thomas Coste et al.

To ensure that large language model (LLM) responses are helpful and non-toxic, a reward model trained on human preference data is usually used. LLM responses with high rewards are then selected through best-of-$n$ (BoN) sampling or the LLM is further optimized to produce responses with high rewards through reinforcement learning from human feedback (RLHF). However, these processes are susceptible to reward overoptimization or `hacking', where responses receive high rewards due to imperfections in the reward model rather than true preference, particularly as prompts or responses deviate from the training data. To address these challenges, we propose to train a Bayesian reward model, which signals higher uncertainty further from the training data distribution. We trained Bayesian reward models using Laplace approximation on LoRA weights, and found that the resulting uncertainty estimates can effectively mitigate reward overoptimization in BoN sampling.

Georgios Papoudakis, Thomas Coste, Zhihao Wu et al.

The utilisation of foundation models as smartphone assistants, termed app agents, is a critical research challenge. These agents aim to execute human instructions on smartphones by interpreting textual instructions and performing actions via the device's interface. While promising, current approaches face significant limitations. Methods that use large proprietary models, such as GPT-4o, are computationally expensive, while those that use smaller fine-tuned models often lack adaptability to out-of-distribution tasks. In this work, we introduce AppVLM, a lightweight Vision-Language Model (VLM). First, we fine-tune it offline on the AndroidControl dataset. Then, we refine its policy by collecting data from the AndroidWorld environment and performing further training iterations. Our results indicate that AppVLM achieves the highest action prediction accuracy in offline evaluation on the AndroidControl dataset, compared to all evaluated baselines, and matches GPT-4o in online task completion success rate in the AndroidWorld environment, while being up to ten times faster. This makes AppVLM a practical and efficient solution for real-world deployment.

Dezhao Luo, Bohan Tang, Kang Li et al.

App agents, which autonomously operate mobile Apps through Graphical User Interfaces (GUIs), have gained significant interest in real-world applications. Yet, they often struggle with long-horizon planning, failing to find the optimal actions for complex tasks with longer steps. To address this, world models are used to predict the next GUI observation based on user actions, enabling more effective agent planning. However, existing world models primarily focus on generating only textual descriptions, lacking essential visual details. To fill this gap, we propose ViMo, the first visual world model designed to generate future App observations as images. For the challenge of generating text in image patches, where even minor pixel errors can distort readability, we decompose GUI generation into graphic and text content generation. We propose a novel data representation, the Symbolic Text Representation~(STR) to overlay text content with symbolic placeholders while preserving graphics. With this design, ViMo employs a STR Predictor to predict future GUIs' graphics and a GUI-text Predictor for generating the corresponding text. Moreover, we deploy ViMo to enhance agent-focused tasks by predicting the outcome of different action options. Experiments show ViMo's ability to generate visually plausible and functionally effective GUIs that enable App agents to make more informed decisions.

Georgios Papoudakis, Thomas Coste, Jianye Hao et al.

Reinforcement learning (RL) using foundation models for policy approximations in multi-turn tasks remains challenging. We identify two main limitations related to sparse reward settings and policy gradient updates, based on which we formulate a key insight: updates from positive samples with high returns typically do not require policy regularisation, whereas updates from negative samples, reflecting undesirable behaviour, can harm model performance. This paper introduces Succeed or Learn Slowly (SoLS), a novel off-policy RL algorithm evaluated on mobile app control tasks. SoLS improves sample efficiency when fine-tuning foundation models for user interface navigation via a modified off-policy actor-critic approach, applying direct policy updates for positive samples and conservative, regularised updates for negative ones to prevent model degradation. We augment SoLS with Successful Transition Replay (STR), which prioritises learning from successful interactions, further improving sample efficiency. We evaluate SoLS on the AndroidWorld benchmark, where it significantly outperforms existing methods (at least 17% relative increase), including prompt-engineering and RL approaches, while requiring substantially fewer computational resources than GPT-4o-based methods with 5-60x faster inference.

Jun Wang, Yu Mao, Yufei Cui et al.

Pre-processing whole slide images (WSIs) can impact classification performance. Our study shows that using fixed hyper-parameters for pre-processing out-of-domain WSIs can significantly degrade performance. Therefore, it is critical to search domain-specific hyper-parameters during inference. However, searching for an optimal parameter set is time-consuming. To overcome this, we propose BAHOP, a novel Similarity-based Basin Hopping optimization for fast parameter tuning to enhance inference performance on out-of-domain data. The proposed BAHOP achieves 5\% to 30\% improvement in accuracy with $\times5$ times faster on average.

Zhongxuan Han, Chaochao Chen, Xiaolin Zheng et al.

Recommender systems are typically biased toward a small group of users, leading to severe unfairness in recommendation performance, i.e., User-Oriented Fairness (UOF) issue. The existing research on UOF is limited and fails to deal with the root cause of the UOF issue: the learning process between advantaged and disadvantaged users is unfair. To tackle this issue, we propose an In-processing User Constrained Dominant Sets (In-UCDS) framework, which is a general framework that can be applied to any backbone recommendation model to achieve user-oriented fairness. We split In-UCDS into two stages, i.e., the UCDS modeling stage and the in-processing training stage. In the UCDS modeling stage, for each disadvantaged user, we extract a constrained dominant set (a user cluster) containing some advantaged users that are similar to it. In the in-processing training stage, we move the representations of disadvantaged users closer to their corresponding cluster by calculating a fairness loss. By combining the fairness loss with the original backbone model loss, we address the UOF issue and maintain the overall recommendation performance simultaneously. Comprehensive experiments on three real-world datasets demonstrate that In-UCDS outperforms the state-of-the-art methods, leading to a fairer model with better overall recommendation performance.

Alexander I. Cowen-Rivers, Daniel Palenicek, Vincent Moens et al.

In this paper, we propose SAMBA, a novel framework for safe reinforcement learning that combines aspects from probabilistic modelling, information theory, and statistics. Our method builds upon PILCO to enable active exploration using novel(semi-)metrics for out-of-sample Gaussian process evaluation optimised through a multi-objective problem that supports conditional-value-at-risk constraints. We evaluate our algorithm on a variety of safe dynamical system benchmarks involving both low and high-dimensional state representations. Our results show orders of magnitude reductions in samples and violations compared to state-of-the-art methods. Lastly, we provide intuition as to the effectiveness of the framework by a detailed analysis of our active metrics and safety constraints.