In Reinforcement Learning (RL), multi-armed Bandit (MAB) problems have found applications across diverse domains such as recommender systems, healthcare, and finance. Traditional MAB algorithms typically assume stationary reward distributions, which limits their effectiveness in real-world scenarios characterized by non-stationary dynamics. This paper addresses this limitation by introducing and evaluating novel Bandit algorithms designed for non-stationary environments. First, we present the Adaptive Discounted Thompson Sampling (ADTS) algorithm, which enhances adaptability through relaxed discounting and sliding window mechanisms to better respond to changes in reward distributions. We then extend this approach to the Portfolio Optimization problem by introducing the Combinatorial Adaptive Discounted Thompson Sampling (CADTS) algorithm, which addresses computational challenges within Combinatorial Bandits and improves dynamic asset allocation. Additionally, we propose a novel architecture called Bandit Networks, which integrates the outputs of ADTS and CADTS, thereby mitigating computational limitations in stock selection. Through extensive experiments using real financial market data, we demonstrate the potential of these algorithms and architectures in adapting to dynamic environments and optimizing decision-making processes. For instance, the proposed bandit network instances present superior performance when compared to classic portfolio optimization approaches, such as capital asset pricing model, equal weights, risk parity, and Markovitz, with the best network presenting an out-of-sample Sharpe Ratio 20% higher than the best performing classical model.

This study introduces CUPID, a novel approach to session-based reciprocal recommendation systems designed for a real-time one-on-one social discovery platform. In such platforms, low latency is critical to enhance user experiences. However, conventional session-based approaches struggle with high latency due to the demands of modeling sequential user behavior for each recommendation process. Additionally, given the reciprocal nature of the platform, where users act as items for each other, training recommendation models on large-scale datasets is computationally prohibitive using conventional methods. To address these challenges, CUPID decouples the time-intensive user session modeling from the real-time user matching process to reduce inference time. Furthermore, CUPID employs a two-phase training strategy that separates the training of embedding and prediction layers, significantly reducing the computational burden by decreasing the number of sequential model inferences by several hundredfold. Extensive experiments on large-scale Azar datasets demonstrate CUPID's effectiveness in a real-world production environment. Notably, CUPID reduces response latency by more than 76% compared to non-asynchronous systems, while significantly improving user engagement.

Mutual funds aim to generate returns above market averages. While predicting their future portfolio allocations can bring economic advantages, the task remains challenging and largely unexplored. To fill that gap, this work frames mutual fund portfolio prediction as a Next Novel Basket Recommendation (NNBR) task, focusing on predicting novel items in a fund's next portfolio. We create a comprehensive benchmark dataset using publicly available data and evaluate the performance of various recommender system models on the NNBR task. Our findings reveal that predicting novel items in mutual fund portfolios is inherently more challenging than predicting the entire portfolio or only repeated items. While state-of-the-art NBR models are outperformed by simple heuristics when considering both novel and repeated items together, autoencoder-based approaches demonstrate superior performance in predicting only new items. The insights gained from this study highlight the importance of considering domain-specific characteristics when applying recommender systems to mutual fund portfolio prediction. The performance gap between predicting the entire portfolio or repeated items and predicting novel items underscores the complexity of the NNBR task in this domain and the need for continued research to develop more robust and adaptable models for this critical financial application.

Cross-domain Recommendation (CDR) systems leverage multi-domain user interactions to improve performance, especially in sparse data or new user scenarios. However, CDR faces challenges such as effectively capturing user preferences and avoiding negative transfer. To address these issues, we propose the Multi-view Disentangled and Adaptive Preference Learning (MDAP) framework. Our MDAP framework uses a multiview encoder to capture diverse user preferences. The framework includes a gated decoder that adaptively combines embeddings from different views to generate a comprehensive user representation. By disentangling representations and allowing adaptive feature selection, our model enhances recommendations' adaptability and effectiveness. Extensive experiments on benchmark datasets demonstrate that our method significantly outperforms state-of-the-art CDR and single-domain models, providing more accurate recommendations and deeper insights into user behavior across different domains.

Multi-task ranking models have become essential for modern real-world recommendation systems. While most recommendation researches focus on designing sophisticated models for specific scenarios, achieving performance improvement for multi-task ranking models across various scenarios still remains a significant challenge. Training all tasks naively can result in inconsistent learning, highlighting the need for the development of multi-task optimization (MTO) methods to tackle this challenge. Conventional methods assume that the optimal joint gradient on shared parameters leads to optimal parameter updates. However, the actual update on model parameters may deviates significantly from gradients when using momentum based optimizers such as Adam, and we design and execute statistical experiments to support the observation. In this paper, we propose a novel Parameter Update Balancing algorithm for multi-task optimization, denoted as PUB. In contrast to traditional MTO method which are based on gradient level tasks fusion or loss level tasks fusion, PUB is the first work to optimize multiple tasks through parameter update balancing. Comprehensive experiments on benchmark multi-task ranking datasets demonstrate that PUB consistently improves several multi-task backbones and achieves state-of-the-art performance. Additionally, experiments on benchmark computer vision datasets show the great potential of PUB in various multi-task learning scenarios. Furthermore, we deployed our method for an industrial evaluation on the real-world commercial platform, HUAWEI AppGallery, where PUB significantly enhances the online multi-task ranking model, efficiently managing the primary traffic of a crucial channel.

This paper studies a game design problem. There are U users and P players. Each player has a set of possible of actions. Each action of a user gives a certain utility to each player. There is a "mediator" that, upon receiving a profile of actions from all the players, will choose which action to display for each user.

This is an excellent paper, proposing a sound idea of approximating a partially defined rating matrix with a combination of multiple low rank matrices of different ranks in order to learn well the head user/item pairs (users and items with lots of ratings) as well as the tail user/item pairs (users and items we few ratings). The idea is introduced clearly. The paper makes a good review of the state-of-the-art, and the experiment section is solid with very convincing results. In reading the introduction, the reader could find controversial the statement in lines 25-27 about the correlation between the number of user-item ratings and the desired rank. One could imagine that a subgroup of users and items have a large number of ratings but in a consistent way, which can be explained with a low rank matrix.

Summary: The authors develop a Gamma-Poisson factorization model that includes metadata and models user preferences and item attractiveness in a dynamic context. They develop a variational inference algorithm and demonstrate that their approach outperforms other methods on five data sets. Quality: The technical quality of this work appears to be sound. For evaluation, the metrics used are in line with the way these systems are actually deployed (e.g., rank-based instead of just RMSE of the ratings). I think the authors sell Gaussian MF a little short.

Choosing the best portable speaker can be a daunting task with the amount of options available today. Whether you're gearing up for a camping trip, a beach outing, or a backyard barbecue, finding the right speaker that delivers on sound quality, durability, and portability is crucial. We've tested dozens of Bluetooth speakers across various price points to help you navigate this crowded market. While many of them sound impressive, comparing them head-to-head allowed us to identify the features that make certain portable wireless speakers stand out. If you're looking primarily for a speaker that works with a voice assistant like Alexa, Google Assistant or Siri, check out our top picks for the best smart speakers. However, for those seeking a versatile portable Bluetooth speaker, we've put together a selection of top performers that cater to a wide range of use cases and preferences. Whether you're after powerful bass, long battery life or rugged design, our recommendations will help you find the best match for your needs. If you're just looking for a small Bluetooth speaker that can kick out some decent volume, the Tribit StormBox Micro 2 fits the bill. The audio quality here is fine; it doesn't stand out in terms of fidelity, but the volume you get from this affordable little speaker is what makes it a good choice. The rubbery rear strap works well on relatively thin things like belts, backpacks and bike handlebars. While it's small and affordable, this mini speaker features a USB-C charging port for your devices in a pinch and you can wirelessly connect two of them for party mode or stereo sound.

The paper changes the model assumptions for recommender systems (RS) to capture phenomena of real world data that has been largely ignored up to now. Instead of assuming a fixed preference over time, the utility of a recommendation is based on the frequency of previous occurrences in a time window w. This captures the fact that humans get bored of repetition. The authors do a great job in showing that this effect occurs in real data. However they still make quite restrictive model assumptions, EDIT{misunderstood this part in the paper, remove comment: i.e. that the utility of an action is only based on the frequency it occurred over the last w times, without taking the positioning into account.