A5026953642- Recommender Systems and Techniques
- Digital Mental Health Interventions
- Mental Health Research Topics
- Behavioral Health and Interventions
- Advanced Bandit Algorithms Research
- Tactile and Sensory Interactions
- Music and Audio Processing
- Interactive and Immersive Displays
- Generative Adversarial Networks and Image Synthesis
- Human Mobility and Location-Based Analysis
- Topic Modeling
- Caching and Content Delivery
- Gaze Tracking and Assistive Technology
- Primate Behavior and Ecology
- Virology and Viral Diseases
- Music Technology and Sound Studies
- Spam and Phishing Detection
- Mobile Crowdsensing and Crowdsourcing
- Adversarial Robustness in Machine Learning
- Aerospace Engineering and Energy Systems
- Neural dynamics and brain function
- Neuroscience and Music Perception
- Impact of Technology on Adolescents
- Functional Brain Connectivity Studies
- Image and Video Quality Assessment
New York University Shanghai
2024
Cornell University
2018-2022
Tsinghua University
2016-2017
Previous work on muscle activity sensing has leveraged specialized sensors such as electromyography and force sensitive resistors. While these show great potential for detecting finger/hand gestures, they require additional hardware that adds to the cost user discomfort. Past research utilized commercial devices, focusing recognizing gross hand gestures. In this we present Serendipity, a new technique unremarkable fine-motor finger gestures using integrated motion (accelerometer gyroscope)...
Schizophrenia is a severe and complex psychiatric disorder with heterogeneous dynamic multi-dimensional symptoms. Behavioral rhythms, such as sleep rhythm, are usually disrupted in people schizophrenia. As such, behavioral rhythm sensing smartphones machine learning can help better understand predict their Our goal to fine-grained symptom changes interpretable models. We computed rhythm-based features from 61 participants 6,132 days of data used multi-task ecological momentary assessment...
Impaired social functioning is a symptom of mental illness (e.g., depression, schizophrenia) and wide range other conditions cognitive decline in the elderly, dementia). Today, assessing relies on subjective evaluations self assessments. We propose different approach collect detailed measures objective mobile sensing data from N=55 outpatients living with schizophrenia to study new methods passively accessing functioning. identify number behavioral patterns data, discuss important...
Modern recommender systems have evolved rapidly along with deep learning models that are well-optimized for overall performance, especially those trained under Empirical Risk Minimization (ERM). However, a recommendation algorithm focuses solely on the average performance may reinforce exposure bias and exacerbate "rich-get-richer" effect, leading to unfair user experience. In simulation study, we demonstrate such gap among various groups is enlarged by an ERM-trained in long-term. To...
Touch interaction on smartwatches suffers from the awkwardness of having to use two hands and "fat finger" problem. We present Float, a wrist-to-finger input approach that enables one-handed touch-free target selection with high efficiency precision using only commercially-available built-in sensors. With user tilts wrist point performs an in-air finger tap click. To realize we first explore appropriate motion space for tilt determine clicking action (finger tap) through user-elicitation...
Implicit feedback (e.g., clicks) is widely used in content recommendations. However, clicks only reflect user preferences according to their first impressions. They do not capture the extent which users continue engage with content. Our analysis shows that more than half of on music and short videos are followed by skips from two real-world datasets. In this paper, we leverage post-click feedback, e.g. completions, improve training evaluation recommenders. Specifically, experiment existing...
The classic Marshmallow Test, where children were offered a choice between one small but immediate reward (eg, marshmallow) or larger two marshmallows) if they waited for period of time, instigated wealth research on the relationships among impulsive responding, self-regulation, and clinical life outcomes. Impulsivity is hallmark feature self-regulation failures that lead to poor health decisions outcomes, making understanding treating impulsivity most important constructs tackle in building...
Background Mobile health technology has demonstrated the ability of smartphone apps and sensors to collect data pertaining patient activity, behavior, cognition. It also offers opportunity understand how everyday passive mobile metrics such as battery life screen time relate mental outcomes through continuous sensing. Impulsivity is an underlying factor in numerous physical problems. However, few studies have been designed help us self-report can improve our understanding impulsive behavior....
Traditionally, recommendation systems are built on the assumption that each service provider has full access to all user data generated its platform. However, with increasing privacy concerns and personal protection regulation, providers such as Google, Twitter, Facebook enabling their users revisit, erase, rectify historical profiles. Future need be robust profile modifications user-controlled filtering. In this paper, we explore how performance may affected by time-sensitive filtering, is,...
Recommender systems play an important role in modern information and e-commerce applications. While increasing research is dedicated to improving the relevance diversity of recommendations, potential risks state-of-the-art recommendation models are under-explored, that is, these could be subject attacks from malicious third parties, through injecting fake user interactions achieve their purposes. This paper revisits adversarially-learned injection attack problem, where injected `behaviors'...
In this paper, we investigate the effects of post-selection feedback for acquiring ultra-small (2-4mm) targets on touchscreens. Post-selection shows contact point touchscreen after a user lifts his/her fingers to increase users' awareness touching. Three experiments are conducted progressively using single crosshair target, two reciprocally acquired and 2D random targets. Results show that in average can reduce touch error rates by 78.4%, with compromise target acquisition time no more than...
Mental illness often emerges during the formative years of adolescence and young adult development interferes with establishment healthy educational, vocational, social foundations. Despite severity symptoms decline in functioning, time between onset receiving appropriate care can be lengthy. A method by which to objectively identify early signs emerging psychiatric could improve intervention strategies. We analyzed a total 405,523 search queries from 105 individuals schizophrenia spectrum...
The last year has been a breakout for podcasts. There are now over 1 million podcast shows and 64 episodes available through public RSS feeds. In the United States, 32% of all people listened to every month, forecasts point global listenership reach 2.2 billion monthly listeners by 2024. workshop on Podcast Recommendations (PodRecs), collocated with RecSys 2020, introduces researchers in other domains recommender systems special characteristics challenges recommendations: how podcasts as...
Recommender Systems are built to retrieve relevant items satisfy users' information needs. The candidate corpus usually consists of a finite set that ready be served, such as videos, products, or articles. With recent advances in Generative AI GPT and Diffusion models, new form recommendation task is yet explored where created by generative models with personalized prompts. Taking image generation an example, single prompt from the user access model, it possible generate hundreds images few...
Selection bias is prevalent in the data for training and evaluating recommendation systems with explicit feedback. For example, users tend to rate items they like. However, when rating an item concerning a specific user, most of algorithms rely too much on his/her (feedback) history. This introduces implicit system, which referred as user feedback-loop this paper. We propose systematic dynamic way correct such obtain more diverse objective recommendations by utilizing temporal information....
High quality user feedback data is essential to training and evaluating a successful music recommendation system, particularly one that has balance the needs of multiple stakeholders. Most existing datasets suffer from noisy self-selection biases inherent in collected by platforms. Using Piki Music dataset 500k ratings over two-year time period, we evaluate performance classic algorithms on three important stakeholders: consumers, well-known artists lesser-known artists. We show matrix...
This tutorial reviews recent developments of deep neural network-based recommendation algorithms and demonstrates how to extend adapt such for diverse application scenarios. The customization is supported by OpenRec framework that modularizes recommenders. consists a lecture two hands-on sessions. It targets intermediate advanced audiences who already possess knowledge networks are interested in applying those the domain recommendation. Materials available at: http://openrec.ai/
Diffusion models excel at generating high-quality images and are easy to extend, making them extremely popular among active users who have created an extensive collection of diffusion with various styles by fine-tuning base such as Stable Diffusion. Recent work has focused on uncovering semantic visual information encoded in components a model, enabling better generation quality more fine-grained control. However, those methods target improving single model overlook the vastly available...
<sec> <title>BACKGROUND</title> Mobile health technology has demonstrated the ability of smartphone apps and sensors to collect data pertaining patient activity, behavior, cognition. It also offers opportunity understand how everyday passive mobile metrics such as battery life screen time relate mental outcomes through continuous sensing. Impulsivity is an underlying factor in numerous physical problems. However, few studies have been designed help us self-report can improve our...
<sec> <title>BACKGROUND</title> The classic Marshmallow Test, where children were offered a choice between one small but immediate reward (eg, marshmallow) or larger two marshmallows) if they waited for period of time, instigated wealth research on the relationships among impulsive responding, self-regulation, and clinical life outcomes. Impulsivity is hallmark feature self-regulation failures that lead to poor health decisions outcomes, making understanding treating impulsivity most...
<sec> <title>UNSTRUCTURED</title> REMOVE </sec>