Sunnyvale, California, United States
Contact Info
736 followers
500+ connections
About
Activity
-
Apple has entered the game! Apple just released a 7B open-source LLM, weights, training code, and dataset! 👀 TL;DR: 🧠 7B base model, trained on…
Apple has entered the game! Apple just released a 7B open-source LLM, weights, training code, and dataset! 👀 TL;DR: 🧠 7B base model, trained on…
Liked by Yafei Wang
-
People often ask why prices like $2.8/m token for Llama 405B, while being super fast, are still profitable at Lepton AI. We've even been asked by a…
People often ask why prices like $2.8/m token for Llama 405B, while being super fast, are still profitable at Lepton AI. We've even been asked by a…
Liked by Yafei Wang
-
With today’s launch of Llama 3.1, not only do we open source the largest most capable OSS LLM with longer context window, and improved performance…
With today’s launch of Llama 3.1, not only do we open source the largest most capable OSS LLM with longer context window, and improved performance…
Liked by Yafei Wang
Experience & Education
Publications
-
How Emotional Support and Informational Support Relate to Linguistic Alignment
Conference on Social Computing, Behavioral-Cultural Modeling & Prediction and Behavior Representation in Modeling and Simulation
Linguistic alignment in text-based communication means that people tend to adjust their language use to one another both in terms of word choice and sentence structure. Previous studies about linguistic alignment suggested that these two forms of adaptation are correlated with each other, and that they build up to alignment at a higher representational level, such as pragmatic alignment for support functions. Two types of social support have been identified as important for online health…
Linguistic alignment in text-based communication means that people tend to adjust their language use to one another both in terms of word choice and sentence structure. Previous studies about linguistic alignment suggested that these two forms of adaptation are correlated with each other, and that they build up to alignment at a higher representational level, such as pragmatic alignment for support functions. Two types of social support have been identified as important for online health communities (OHCs): emotional and informational support between support seekers and support providers. Do the two lower-level alignment measures (lexical and syntactic) relate to these two types of social support in the same way or, are they different? Our hypothesis was that they are similar, due to their correlation relationship. However, we found that, based on an analysis of a 10-year online forum for cancer survivors, the lower-level alignment measures have distinct relationships to the two higher-level support functions. In this paper, we describe this finding and its implications regarding potential refinement of the Interactive Alignment Model.
Other authors -
-
What do people like to “share” about obesity? A content analysis of frequent retweets about obesity on Twitter
Health Commun. 2015 Jun 18:1-14
Twitter has been recognized as a useful channel for the sharing and dissemination of health information, owing in part to its “retweet” function. This study reports findings from a content analysis of frequently retweeted obesity-related tweets to identify the prevalent beliefs and attitudes about obesity on Twitter, as well as key message features that prompt retweeting behavior conducive to maximizing the reach of health messages on Twitter. The findings show that tweets that are emotionally…
Twitter has been recognized as a useful channel for the sharing and dissemination of health information, owing in part to its “retweet” function. This study reports findings from a content analysis of frequently retweeted obesity-related tweets to identify the prevalent beliefs and attitudes about obesity on Twitter, as well as key message features that prompt retweeting behavior conducive to maximizing the reach of health messages on Twitter. The findings show that tweets that are emotionally evocative, humorous, and concern individual-level causes for obesity were more frequently retweeted than their counterparts. Specifically, tweets that evoke amusement were retweeted most frequently, followed by tweets evoking contentment, surprise, and anger. In regard to humor, derogatory jokes were more frequently retweeted than nonderogatory ones, and in terms of specific types of humor, weight-related puns, repartee, and parody were shared frequently. Consistent with extant literature about obesity, the findings demonstrated the predominance of the individual-level (e.g., problematic diet, lack of exercise) over social-level causes for obesity (e.g., availability of cheap and unhealthy food). Implications for designing social-media-based health campaign messages are discussed.
Other authors -
-
Linguistic Adaptation in Conversation Threads: Analyzing Alignment in Online Health Communities
Cognitive Modeling and Computational Linguistics, ACL, 2014
This paper quantifies linguistic alignment in long-term, online forum conversations; we demonstrate the existence of alignment at lexical and syntactic levels. Topicality- and priming-induced decay effects exist at a large scale in forum threads. However, adaptation to messages of specific social and interactional roles suggests that initial posts obtain a special role in the discourse.
Other authors -
-
Understanding topics, sentiment, and influence in an online cancer survivor community
J Natl Cancer Inst Monogr
Online cancer communities help members support one another, provide new perspectives about living with cancer, normalize experiences, and reduce isolation. The American Cancer Society's 166000-member Cancer Survivors Network (CSN) is the largest online peer support community for cancer patients, survivors, and caregivers. Sentiment analysis and topic modeling were applied to CSN breast and colorectal cancer discussion posts from 2005 to 2010 to examine how sentiment change of thread initiators,…
Online cancer communities help members support one another, provide new perspectives about living with cancer, normalize experiences, and reduce isolation. The American Cancer Society's 166000-member Cancer Survivors Network (CSN) is the largest online peer support community for cancer patients, survivors, and caregivers. Sentiment analysis and topic modeling were applied to CSN breast and colorectal cancer discussion posts from 2005 to 2010 to examine how sentiment change of thread initiators, a measure of social support, varies by discussion topic. The support provided in CSN is highest for medical, lifestyle, and treatment issues. Threads related to 1) treatments and side effects, surgery, mastectomy and reconstruction, and decision making for breast cancer, 2) lung scans, and 3) treatment drugs in colon cancer initiate with high negative sentiment and produce high average sentiment change. Using text mining tools to assess sentiment, sentiment change, and thread topics provides new insights that community managers can use to facilitate member interactions and enhance support outcomes.
Other authorsSee publication -
All-Visible-k-Nearest-Neighbor Queries
DEXA. Springer Berlin Heidelberg, 2012
The All-k-Nearest-Neighbor (AkNN) operation is common in many applications such as GIS and data analysis/mining. In this paper, for the first time, we study a novel variant of AkNN queries, namely All-Visible-k-Nearest-Neighbor (AVkNN) query, which takes into account the impact of obstacles on the visibility of objects. Given a data set P, a query set Q, and an obstacle set O, an AVkNN query retrieves for each point/object in Q its visible k nearest neighbors in P. We formalize the AVkNN query,…
The All-k-Nearest-Neighbor (AkNN) operation is common in many applications such as GIS and data analysis/mining. In this paper, for the first time, we study a novel variant of AkNN queries, namely All-Visible-k-Nearest-Neighbor (AVkNN) query, which takes into account the impact of obstacles on the visibility of objects. Given a data set P, a query set Q, and an obstacle set O, an AVkNN query retrieves for each point/object in Q its visible k nearest neighbors in P. We formalize the AVkNN query, and then propose efficient algorithms for AVkNN retrieval, assuming that P, Q, and O are indexed by conventional data-partitioning indexes (e.g., R-trees). Our approaches employ pruning techniques and introduce a new pruning metric called VMDIST. Extensive experiments using both real and synthetic datasets demonstrate the effectiveness of our presented pruning techniques and the performance of our proposed algorithms.
Other authors -
-
Pragmatic Alignment on Social Support Type in Health Forum Conversations
Cognitive Modeling and Computational Linguistics, NAACL, 2015
We study whether participants show alignment on social support in an online community. While adaptation can occur at linguistic and semantic levels, the relationships between alignment at multiple levels are neither theoretically nor empirically well understood. Our results indicate pragmatic alignment of forum participants along the axis of support type. We also find that lexical alignment is correlated with emotional support, and that the amount of lexical alignment is also correlated with…
We study whether participants show alignment on social support in an online community. While adaptation can occur at linguistic and semantic levels, the relationships between alignment at multiple levels are neither theoretically nor empirically well understood. Our results indicate pragmatic alignment of forum participants along the axis of support type. We also find that lexical alignment is correlated with emotional support, and that the amount of lexical alignment is also correlated with the amount of pragmatic alignment. This finding can contribute to improving our understanding about the linguistic signature of different types of support, and enhancing theoretical model about the Interactive Alignment Model in a multi-party peer support conversation context.
Other authors -
-
Quantified Self Meets Social Media: Sharing of Weight Updates on Twitter
Proceedings of the 6th International Conference on Digital Health Conference, ACM
An increasing number of people use wearables and other smart devices to quantify various health conditions, ranging from sleep patterns, to body weight, to heart rates. Of these "Quantified Selfs", many choose to openly share their data via online social networks such as Twitter and Facebook. In this study, we use data for users who have chosen to connect their smart scales to Twitter, providing both a reliable time series of their body weight, as well as insights into their social…
An increasing number of people use wearables and other smart devices to quantify various health conditions, ranging from sleep patterns, to body weight, to heart rates. Of these "Quantified Selfs", many choose to openly share their data via online social networks such as Twitter and Facebook. In this study, we use data for users who have chosen to connect their smart scales to Twitter, providing both a reliable time series of their body weight, as well as insights into their social surroundings and general online behavior. Concretely, we look at which social media features are predictive of physical sta- tus, such as body weight at the individual level, and activity patterns at the population level. We show that it is possible to predict an individual’s weight using their online social behaviors, such as their self-description and tweets. Weekly and monthly patterns of quantified-self behaviors are also discovered. These findings could contribute to building mod- els to monitor public health and to have more customized personal training interventions.
While there are many studies using either quantified self or social media data in isolation, this is one of the few that combines the two data sources and, to the best of our knowl- edge, the only one that uses public data.Other authors -
Courses
-
Advanced Data Structure & Algorithm Analysis
-
-
Advanced Social Media Management
IST 597I
-
Analysis of Discrete Data
STAT 504
-
Applied Statistics
STAT 500
-
Artificial Intelligence
IST 597F
-
Assembly Language
-
-
C Programming Language
-
-
Computational Graphics
-
-
Computer Architecture
-
-
Computer Networks
-
-
Computer Organization
-
-
Database Management System Design
-
-
Discrete Mathematics
-
-
Java Programming
-
-
Machine Learning
CSE 598A
-
Object-Oriented Programming
-
-
Pattern Recognition
CSE 583
-
Principles of Operating Systems
-
-
Regression Methods
STAT 501
-
Software Engineering
-
-
Web Analytics
IST 597J
Projects
-
Weight loss prediction using quantified self data on Twitter Data
-
Size Constrained Community Detection in Medical Forums
Developed a size-constrained user group detection algorithm based on users’ interests and interactions for online communities with Python.
Other creators -
-
Target dependent subjectivity classification on Twitter Data
Developed a semi-supervised infrastructure predicting target-dependent subjectivity on twitter data using Word2Vec, implemented with Java.
-
Topic Modeling on Cancer Survivor Network
Languages
-
English
Full professional proficiency
-
Mandarin
Native or bilingual proficiency
More activity by Yafei
-
Shuffling node features boosts GNN performance! 🤯 Exploring this *counterintuitive* observation helps us understand when and how GNNs are…
Shuffling node features boosts GNN performance! 🤯 Exploring this *counterintuitive* observation helps us understand when and how GNNs are…
Liked by Yafei Wang
-
Our team builds ML models and optimization systems that have high impact for both Uber's Mobility and Delivery businesses. We're hiring for multiple…
Our team builds ML models and optimization systems that have high impact for both Uber's Mobility and Delivery businesses. We're hiring for multiple…
Liked by Yafei Wang
-
An interesting blog to check out if you are interested in how we retrieve high quality candidates in a large scale graph recommendation system at…
An interesting blog to check out if you are interested in how we retrieve high quality candidates in a large scale graph recommendation system at…
Shared by Yafei Wang
-
We wrote a new blog about candidate generation! We dive into the “People You May Know” (PYMK) system to discuss dealing with potentially billions of…
We wrote a new blog about candidate generation! We dive into the “People You May Know” (PYMK) system to discuss dealing with potentially billions of…
Liked by Yafei Wang
-
iPad Air (13 inch) + reMarkable 2 is my recent favorite tool
iPad Air (13 inch) + reMarkable 2 is my recent favorite tool
Liked by Yafei Wang
-
We are excited to announce our offering in Databricks. Kumo.AI on Databricks lets you; 1. Deliver More Predictions: Build your graph once by…
We are excited to announce our offering in Databricks. Kumo.AI on Databricks lets you; 1. Deliver More Predictions: Build your graph once by…
Liked by Yafei Wang
-
I am thrilled to share that our team at AI Algorithms Foundation (AIAF) has achieved a significant milestone: 4 papers accepted to KDD 2024! 🎉📚…
I am thrilled to share that our team at AI Algorithms Foundation (AIAF) has achieved a significant milestone: 4 papers accepted to KDD 2024! 🎉📚…
Liked by Yafei Wang
-
We just posted an update to our preprint titled "Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context". To highlight…
We just posted an update to our preprint titled "Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context". To highlight…
Liked by Yafei Wang
-
Today is my first day at LinkedIn. I'm very happy to join the AI growth team. I love the culture and what LinkedIn represents: connecting the world's…
Today is my first day at LinkedIn. I'm very happy to join the AI growth team. I love the culture and what LinkedIn represents: connecting the world's…
Liked by Yafei Wang
-
Hello Network, my team is hiring multiple entry or mid-level Applied Scientists for Rufus, Amazon's next-generation AI-driven search and shopping…
Hello Network, my team is hiring multiple entry or mid-level Applied Scientists for Rufus, Amazon's next-generation AI-driven search and shopping…
Liked by Yafei Wang
-
Career update: Delighted to share that I will be joining University of Southern California as an Assistant Professor of Aerospace and Mechanical…
Career update: Delighted to share that I will be joining University of Southern California as an Assistant Professor of Aerospace and Mechanical…
Liked by Yafei Wang
-
I wrote about CPU vs. GPU for neural networks: https://lnkd.in/gJzTm6Vt
I wrote about CPU vs. GPU for neural networks: https://lnkd.in/gJzTm6Vt
Liked by Yafei Wang
-
Recently I presented at an internal reading group for a deep dive into recent advances in generative LLM modeling. All materials are from open-source…
Recently I presented at an internal reading group for a deep dive into recent advances in generative LLM modeling. All materials are from open-source…
Liked by Yafei Wang
-
Just had an incredible experience with Kumo.AI over the weekend! 🚀 A bit of context: I joined Kumo's first hackathon while juggling work-from-home…
Just had an incredible experience with Kumo.AI over the weekend! 🚀 A bit of context: I joined Kumo's first hackathon while juggling work-from-home…
Liked by Yafei Wang
Other similar profiles
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore MoreOthers named Yafei Wang in United States
-
Yafei Wang
-
Yafei Wang
-
Yafei Wang
Financial Analyst at Financial Data Technologies, Ltd. (FDT)
-
Yafei Wang
California State University-Los Angeles Student
17 others named Yafei Wang in United States are on LinkedIn
See others named Yafei Wang