We introduce Dynamic Contextual Markov Decision Processes (DCMDPs), a novel reinforcement learning framework for history-dependent environments that generalizes the contextual MDP framework to handle non-Markov environments, where contexts change over time. We consider special cases of the model, with a focus on logistic DCMDPs, which break the exponential dependence on history length by leveraging aggregation functions to determine context transitions. This special structure allows us to derive an upper-confidence-bound style algorithm for which we establish regret bounds. Motivated by our theoretical results, we introduce a practical model-based algorithm for logistic DCMDPs that plans in a latent space and uses optimism over history-dependent features. We demonstrate the efficacy of our approach on a recommendation task (using MovieLens data) where user behavior dynamics evolve in response to recommendations.

Off-Policy reinforcement learning has been a driving force for the state-of-the-art conversational AIs leading to more natural humanagent interactions and improving the user satisfaction for goal-oriented agents. However, in large-scale commercial settings, it is often challenging to balance between policy improvements and experience continuity on the broad spectrum of applications handled by such system. In the literature, off-policy evaluation and guard-railing on aggregate statistics has been commonly used to address this problem. In this paper, we propose a method for curating and leveraging high-precision samples sourced from historical regression incident reports to validate, safe-guard, and improve policies prior to the online deployment. We conducted extensive experiments using data from a real-world conversational system and actual regression incidents. The proposed method is currently deployed in our production system to protect customers against broken experiences and enable long-term policy improvements.

Building recommendation systems via federated learning (FL) is a new emerging challenge for advancing next-generation Internet service and privacy protection. Existing approaches train shared item embedding by FL while keeping the user embedding private on client side. However, item embedding identical for all clients cannot capture users' individual differences on perceiving the same item and thus leads to poor personalization. Moreover, dense item embedding in FL results in expensive communication cost and latency. To address these challenges, we propose Federated Recommendation with Additive Personalization (FedRAP), which learns a global view of items via FL and a personalized view locally on each user. FedRAP enforces sparsity of the global view to save FL's communication cost and encourages difference between the two views through regularization. We propose an effective curriculum to learn the local and global views progressively with increasing regularization weights. To produce recommendations for an user, FedRAP adds the two views together to obtain a personalized item embedding. FedRAP achieves the best performance in FL setting on multiple benchmarks. It outperforms recent federated recommendation methods and several ablation study baselines.

When we first reviewed Apple's HomePod mini in 2020, we had some reservations. While it was a much better value than the original HomePod, it still had some of the same limitations of its bigger sibling. Siri wasn't as bright as Alexa or Google Assistant, the HomeKit ecosystem was limited and there were no real alternatives to Apple Music for on-demand tunes. You bought a mini for Apple's tight integration, and not much else. Apple has significantly expanded the HomePod mini's functionality.

In the current landscape of ever-increasing levels of digitalization, we are facing major challenges pertaining to scalability. Recommender systems have become irreplaceable both for helping users navigate the increasing amounts of data and, conversely, aiding providers in marketing products to interested users. The growing awareness of discrimination in machine learning methods has recently motivated both academia and industry to research how fairness can be ensured in recommender systems. For recommender systems, such issues are well exemplified by occupation recommendation, where biases in historical data may lead to recommender systems relating one gender to lower wages or to the propagation of stereotypes. In particular, consumer-side fairness, which focuses on mitigating discrimination experienced by users of recommender systems, has seen a vast number of diverse approaches for addressing different types of discrimination. The nature of said discrimination depends on the setting and the applied fairness interpretation, of which there are many variations. This survey serves as a systematic overview and discussion of the current research on consumer-side fairness in recommender systems. To that end, a novel taxonomy based on high-level fairness interpretation is proposed and used to categorize the research and their proposed fairness evaluation metrics. Finally, we highlight some suggestions for the future direction of the field.

Recommending points of interest (POIs) is a challenging task that requires extracting comprehensive location data from location-based social media platforms. To provide effective location-based recommendations, it's important to analyze users' historical behavior and preferences. In this study, we present a sophisticated location-aware recommendation system that uses Bidirectional Encoder Representations from Transformers (BERT) to offer personalized location-based suggestions. Our model combines location information and user preferences to provide more relevant recommendations compared to models that predict the next POI in a sequence. Our experiments on two benchmark dataset show that our BERT-based model outperforms various state-of-the-art sequential models. Moreover, we see the effectiveness of the proposed model for quality through additional experiments.

A Conversational Recommender System (CRS) offers increased transparency and control to users by enabling them to engage with the system through a real-time multi-turn dialogue. Recently, Large Language Models (LLMs) have exhibited an unprecedented ability to converse naturally and incorporate world knowledge and common-sense reasoning into language understanding, unlocking the potential of this paradigm. However, effectively leveraging LLMs within a CRS introduces new technical challenges, including properly understanding and controlling a complex conversation and retrieving from external sources of information. These issues are exacerbated by a large, evolving item corpus and a lack of conversational data for training. In this paper, we provide a roadmap for building an end-to-end large-scale CRS using LLMs. In particular, we propose new implementations for user preference understanding, flexible dialogue management and explainable recommendations as part of an integrated architecture powered by LLMs. For improved personalization, we describe how an LLM can consume interpretable natural language user profiles and use them to modulate session-level context. To overcome conversational data limitations in the absence of an existing production CRS, we propose techniques for building a controllable LLM-based user simulator to generate synthetic conversations. As a proof of concept we introduce RecLLM, a large-scale CRS for YouTube videos built on LaMDA, and demonstrate its fluency and diverse functionality through some illustrative example conversations.

In many scenarios, recommender system user interaction data such as clicks or ratings is sparse, and item turnover rates (e.g., new articles, job postings) high. Given this, the integration of contextual "side" information in addition to user-item ratings is highly desirable. Whilst there are algorithms that can handle both rating and contextual data simultaneously, these algorithms are typically limited to making only in-sample recommendations, suffer from the curse of dimensionality, and do not incorporate multi-armed bandit (MAB) policies for long-term cumulative reward optimization. We propose multi-view interactive topic regression (MV-ICTR) a novel partially online latent factor recommender algorithm that incorporates both rating and contextual information to model item-specific feature dependencies and users' personal preferences simultaneously, with multi-armed bandit policies for continued online personalization. The result is significantly increased performance on datasets with high percentages of cold-start users and items.

The cold start problem is a common challenge in various domains, including media use cases such as predicting viewership for newly launched shows on Over-The-Top (OTT) platforms. In this study, we propose a generic approach to tackle cold start problems by leveraging metadata and employing multi-model ensemble techniques. Our methodology includes feature engineering, model selection, and an ensemble approach based on a weighted average of predictions. The performance of our proposed method is evaluated using various performance metrics. Our results indicate that the multi-model ensemble approach significantly improves prediction accuracy compared to individual models.

Deep learning-based recommendation systems (e.g., DLRMs) are widely used AI models to provide high-quality personalized recommendations. Training data used for modern recommendation systems commonly includes categorical features taking on tens-of-millions of possible distinct values. These categorical tokens are typically assigned learned vector representations, that are stored in large embedding tables, on the order of 100s of GB. Storing and accessing these tables represent a substantial burden in commercial deployments. Our work proposes MEM-REC, a novel alternative representation approach for embedding tables. MEM-REC leverages bloom filters and hashing methods to encode categorical features using two cache-friendly embedding tables. The first table (token embedding) contains raw embeddings (i.e. learned vector representation), and the second table (weight embedding), which is much smaller, contains weights to scale these raw embeddings to provide better discriminative capability to each data point. We provide a detailed architecture, design and analysis of MEM-REC addressing trade-offs in accuracy and computation requirements, in comparison with state-of-the-art techniques. We show that MEM-REC can not only maintain the recommendation quality and significantly reduce the memory footprint for commercial scale recommendation models but can also improve the embedding latency. In particular, based on our results, MEM-REC compresses the MLPerf CriteoTB benchmark DLRM model size by 2900x and performs up to 3.4x faster embeddings while achieving the same AUC as that of the full uncompressed model.