How Intention Informed Recommendations Modulate Choices: A Field Study of Spoken Word Content

Yang, Longqi; Sobolev, Michael; Wang, Yu; Chen, Jenny; Dunne, Drew; Tsangouri, Christina; Dell, Nicola; Нааман, Мор; Estrin, Deborah

doi:10.1145/3308558.3313540

Cited by 14 publications

(13 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, improvements in machine learning [39], increased availability of relevant data [9], an enhanced battery and network performance mean that efficacious context models are increasingly practicable [11]. Although research has attempted to encourage behaviour change through recommendations [18,20,21,38,42,49], machine learning models [28,41] and timely interventions [30,32,39], CoCo is the first system, that we know of, which combines alcohol, caffeine and cortisol sensors with a functional context model in order to encourage specific user behaviours.…”

Section: Background and Related Workmentioning

confidence: 99%

Context-Aware Wearables

Snooks

Lindley

Richards

et al. 2021

Extended Abstracts of the 2021 CHI Conference on Human Factors in Computing Systems

View full text Add to dashboard Cite

We present Connected Companion (CoCo), a health tracking wearable that provides users with timely, context-relevant notifications aimed at improving wellness. Traditionally, self-tracking wearables report basic health data such as resting heart rate; these data are visualised and positive behaviours (e.g. exercising often) are encouraged with rudimentary gamification (e.g. award badges) and notification systems. CoCo is the first wearable to combine caffeine, alcohol and cortisol sensors, a context network (which predicts user context), and a wellness model (which establishes per-user wellness measures). Working in tandem these provide users with notifications that encourage discrete behaviours intended to optimise user-wellness per very specific biological and social contexts. The paper describes the (sometimes unexpected) results of a user-study intended to evaluate CoCo's efficacy and we conclude with a discussion about the power and responsibility that comes with attempts to build context-aware computing systems. CCS CONCEPTS • Human-centered computing à User interface design; Information visualization; • Applied computing à Health informatics; • Computer systems organization à Sensors and actuators; • General and reference à Design.

show abstract

Section: Background and Related Workmentioning

confidence: 99%

Context-Aware Wearables

Snooks

Lindley

Richards

et al. 2021

Extended Abstracts of the 2021 CHI Conference on Human Factors in Computing Systems

View full text Add to dashboard Cite

show abstract

“…While podcast consumption has grown rapidly in recent years, there is comparatively little research on podcast recommendation. One of the few papers is recent work by Yang et al, which compares the effects of intention-informed recommendations with classic intention-agnostic systems and shows that a recommender can boost a user's aspiration-based consumption [25].…”

Section: Podcast Recommendationsmentioning

confidence: 99%

“…• Age bucket: We are interested in evaluating performance across all age ranges in order to see if some age ranges were being underserved by the model. Figure 6 shows that our model performs best in the [25][26][27][28][29] age bucket with a 50% improvement over the baseline, but performs well across all age buckets. • Gender Finally, our model's performance for self-reported genders are shown in Figure 7.…”

Section: Model Performance In User Cohortsmentioning

confidence: 99%

Recommending Podcasts for Cold-Start Users Based on Music Listening and Taste

Nazari

Charbuillet

Pages

et al. 2020

Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval

View full text Add to dashboard Cite

Recommender systems are increasingly used to predict and serve content that aligns with user taste, yet the task of matching new users with relevant content remains a challenge. We consider podcasting to be an emerging medium with rapid growth in adoption, and discuss challenges that arise when applying traditional recommendation approaches to address the cold-start problem. Using music consumption behavior, we examine two main techniques in inferring Spotify users preferences over more than 200k podcasts. Our results show significant improvements in consumption of up to 50% for both offline and online experiments. We provide extensive analysis on model performance and examine the degree to which music data as an input source introduces bias in recommendations.

show abstract

“…This perspective brings user agency into the center, prioritizing the the ability for models to be as adaptable as they are accurate, able to accommodate arbitrary changes in the interests of individuals. Studies nd positive e ects of allowing users to exert greater control in recommendation systems [21,51]. While there are many system-level or post-hoc approaches to incorporating user feedback, we focus directly on the machine learning model that powers recommendations.…”

Section: Introductionmentioning

confidence: 99%

Recommendations and User Agency: The Reachability of Collaboratively-Filtered Information

Dean,

Rich,

Recht

2019

Preprint

View full text Add to dashboard Cite

Recommender systems often rely on models which are trained to maximize accuracy in predicting user preferences. When the systems are deployed, these models determine the availability of content and information to di erent users. The gap between these objectives gives rise to a potential for unintended consequences, contributing to phenomena such as lter bubbles and polarization. In this work, we consider directly the information availability problem through the lens of user recourse. Using ideas of reachability, we propose a computationally e cient audit for top-N linear recommender models. Furthermore, we describe the relationship between model complexity and the e ort necessary for users to exert control over their recommendations. We use this insight to provide a novel perspective on the user cold-start problem. Finally, we demonstrate these concepts with an empirical investigation of a state-of-the-art model trained on a widely used movie ratings dataset.

show abstract

How Intention Informed Recommendations Modulate Choices: A Field Study of Spoken Word Content

Cited by 14 publications

References 40 publications

Context-Aware Wearables

Context-Aware Wearables

Recommending Podcasts for Cold-Start Users Based on Music Listening and Taste

Recommendations and User Agency: The Reachability of Collaboratively-Filtered Information

Contact Info

Product

Resources

About