The rapid progress of foundation models has led to the prosperity of autonomous agents, which leverage the universal capabilities of foundation models to conduct reasoning, decision-making, and environmental interaction. However, the efficacy of agents remains limited when operating in intricate, realistic environments. In this work, we introduce the principles of $\mathbf{U}$nified $\mathbf{A}$lignment for $\mathbf{A}$gents ($\mathbf{UA}^2$), which advocate for the simultaneous alignment of agents with human intentions, environmental dynamics, and self-constraints such as the limitation of monetary budgets. From the perspective of $\mathbf{UA}^2$, we review the current agent research and highlight the neglected factors in existing agent benchmarks and method candidates. We also conduct proof-of-concept studies by introducing realistic features to WebShop, including user profiles to demonstrate intentions, personalized reranking for complex environmental dynamics, and runtime cost statistics to reflect self-constraints. We then follow the principles of $\mathbf{UA}^2$ to propose an initial design of our agent, and benchmark its performance with several candidate baselines in the retrofitted WebShop. The extensive experimental results further prove the importance of the principles of $\mathbf{UA}^2$. Our research sheds light on the next steps of autonomous agent research with improved general problem-solving abilities.

State-of-the-art rule-based and classification-based food recommendation systems face significant challenges in becoming practical and useful. This difficulty arises primarily because most machine learning models struggle with problems characterized by an almost infinite number of classes and a limited number of samples within an unbalanced dataset. Conversely, the emergence of Large Language Models (LLMs) as recommendation engines offers a promising avenue. However, a general-purpose Recommendation as Language Processing (RLP) approach lacks the critical components necessary for effective food recommendations. To address this gap, we introduce Food Recommendation as Language Processing (F-RLP), a novel framework that offers a food-specific, tailored infrastructure. F-RLP leverages the capabilities of LLMs to maximize their potential, thereby paving the way for more accurate, personalized food recommendations.

Finding love in the modern era seems to be largely reliant on dating apps like Tinder and Bumble. In fact, today's young adults – who were born just as online dating really started to take over – wouldn't know it any other way. But for hundreds of years, the job of matchmaker largely fell to the internet's ink-and-paper predecessor – the local newspaper. In the run-up to Valentine's Day, hilarious, sad and often poignant clippings reveal how people found a romantic match in times past. Ranging from articles to lovestruck poems and lonely hearts columns, they might provide an unlikely source of inspiration if you've someone to woo.

Applied recommender systems research is in a curious position. While there is a very rigorous protocol for measuring performance by A/B testing, best practice for finding a `B' to test does not explicitly target performance but rather targets a proxy measure. The success or failure of a given A/B test then depends entirely on if the proposed proxy is better correlated to performance than the previous proxy. No principle exists to identify if one proxy is better than another offline, leaving the practitioners shooting in the dark. The purpose of this position paper is to question this anti-Utopian thinking and argue that a non-standard use of the deep learning stacks actually has the potential to unlock reward optimizing recommendation.

The ability to shape health behaviors of large populations automatically, across wearable types and disease conditions at scale has tremendous potential to improve global health outcomes. We designed and implemented an AI driven platform for digital algorithmic nudging, enabled by a Graph-Neural Network (GNN) based Recommendation System, and granular health behavior data from wearable fitness devices. Here we describe the efficacy results of this platform with its capabilities of personalized and contextual nudging to $n=84,764$ individuals over a 12-week period in Singapore. We statistically validated that participants in the target group who received such AI optimized daily nudges increased daily physical activity like step count by 6.17% ($p = 3.09\times10^{-4}$) and weekly minutes of Moderate to Vigorous Physical Activity (MVPA) by 7.61% ($p = 1.16\times10^{-2}$), compared to matched participants in control group who did not receive any nudges. Further, such nudges were very well received, with a 13.1% of nudges sent being opened (open rate), and 11.7% of the opened nudges rated useful compared to 1.9% rated as not useful thereby demonstrating significant improvement in population level engagement metrics.

Collaborator recommendation is an important task in academic domain. Most of the existing approaches have the assumption that the recommendation system only need to recommend a specific researcher for the task. However, academic successes can be owed to productive collaboration of a whole academic team. In this work, we propose a new task: academic team worker recommendation: with a given status: student, assistant professor or prime professor, research interests and specific task, we can recommend an academic team formed as (prime professor, assistant professor, student). For this task, we propose a model CQBG-R(Citation-Query Blended Graph-Ranking). The key ideas is to combine the context of the query and the papers with the graph topology to form a new graph(CQBG), which can target at the research interests and the specific research task for this time. The experiment results show the effectiveness of the proposed method.

Fine-tuning is becoming widely used for leveraging the power of pre-trained foundation models in new downstream tasks. While there are many successes of fine-tuning on various tasks, recent studies have observed challenges in the generalization of fine-tuned models to unseen distributions (i.e., out-of-distribution; OOD). To improve OOD generalization, some previous studies identify the limitations of fine-tuning data and regulate fine-tuning to preserve the general representation learned from pre-training data. However, potential limitations in the pre-training data and models are often ignored. In this paper, we contend that overly relying on the pre-trained representation may hinder fine-tuning from learning essential representations for downstream tasks and thus hurt its OOD generalization. It can be especially catastrophic when new tasks are from different (sub)domains compared to pre-training data. To address the issues in both pre-training and fine-tuning data, we propose a novel generalizable fine-tuning method LEVI, where the pre-trained model is adaptively ensembled layer-wise with a small task-specific model, while preserving training and inference efficiencies. By combining two complementing models, LEVI effectively suppresses problematic features in both the fine-tuning data and pre-trained model and preserves useful features for new tasks. Broad experiments with large language and vision models show that LEVI greatly improves fine-tuning generalization via emphasizing different views from fine-tuning data and pre-trained features.

Spoken dialogue systems have transformed human-machine interaction by providing real-time responses to queries. However, misunderstandings between the user and system persist. This study explores the significance of interactional language in dialogue repair between virtual assistants and users by analyzing interactions with Google Assistant and Siri, focusing on their utilization and response to the other-initiated repair strategy "huh?" prevalent in human-human interaction. Findings reveal several assistant-generated strategies but an inability to replicate human-like repair strategies such as "huh?". English and Spanish user acceptability surveys show differences in users' repair strategy preferences and assistant usage, with both similarities and disparities among the two surveyed languages. These results shed light on inequalities between interactional language in human-human interaction and human-machine interaction, underscoring the need for further research on the impact of interactional language in human-machine interaction in English and beyond.

Ryan and Randy met at a sex party in 2019 and started dating shortly after. By month four, they made the relationship official, eventually moved into a two-story house in Los Angeles together, and did all the things happy couples do: date nights, vacation with friends, support one another's ambitions. Then, in 2022, they decided to open the relationship. As Covid-19 restrictions loosened, "we were being exposed to other attractions and to other people who were seeking our attention," Ryan says. "We both knew we had attractions to other people. It was, let's talk about being open and see what that means for us. Because being open can mean different things to different people."

If you're already onboard with Alexa and have decided you want a smart display, a new deal on Amazon's latest Echo Show 8 may be of interest. The 8-inch display is currently down to 90 at Amazon, Target, Best Buy and other retailers, which matches the lowest price we've seen since the device was unveiled last September. Amazon normally sells the smart display for 150, though we saw it drop to 105 for much of the holiday season. Amazon's offer also includes a Sengled color smart bulb for no extra cost. That bulb is compatible with the Matter smart home standard, and we recommend a similar model in our guide to the best smart lights. This matches the best price we've tracked for the latest iteration of Amazon's 8-inch smart display and includes a smart bulb for no extra cost.