[1]
Access: A Mobile Application to Improve Accessibility
Video Showcase Presentations
/
Yang, Yi
/
Hu, Yunqi
/
Hong, Yidi
/
Joshi, Varun
/
Kolathumani, Radhika
Extended Abstracts of the ACM CHI'16 Conference on Human Factors in
Computing Systems
2016-05-07
v.2
p.5
© Copyright 2016 ACM
Summary: This video introduces Access, a mobile application to provide information
about accessibility information of local establishment and public spaces. We
partnered with a local nonprofit in Jackson, MI which allowed us to access
existing data they had collected in the three county areas of Jackson,
Hillsdale and Lenawee in Michigan. The application is focused on providing
information on accessibility of local establishments and public spaces, and how
the mechanism works will be clearly explained in the video. Additionally, the
video demonstrates details of user interfaces of Access and information flow,
and also provides user scenarios depicting how the solution fits in the life of
wheelchair users and solves problems.
[2]
Video driven pedestrian visualization with characteristic appearances
Crowds & human
/
Hu, Yu
/
Wu, Wei
/
Zhou, Zhong
Proceedings of the 2015 ACM Symposium on Virtual Reality Software and
Technology
2015-11-13
p.183-186
© Copyright 2015 ACM
Summary: Augmented virtual environment (AVE) could visualize plausible live views
from videos by projecting dynamic imagery to the 3D environment. Static objects
in a video can be rendered in new views since they are easily modeled
beforehand, while moving ones that don't have exact online models will be
distorted from different views without proper depth. To cope with the problem,
we introduce a novel method to visualize pedestrians, which are common in
outdoor surveillance. Our method detects pedestrians and produces their
trajectories. Then such pedestrian characteristic appearances as geometric
information, texture and walking animation are transferred to a stand-in 3D
animation model in the virtual environment. Experiments show our visualization
can reveal person's characteristic appearances.
[3]
Collaborative Fashion Recommendation: A Functional Tensor Factorization
Approach
Session 3: Emotional and Social Signals in Multimedia
/
Hu, Yang
/
Yi, Xi
/
Davis, Larry S.
Proceedings of the 2015 ACM International Conference on Multimedia
2015-10-26
p.129-138
© Copyright 2015 ACM
Summary: With the rapid expansion of online shopping for fashion products, effective
fashion recommendation has become an increasingly important problem. In this
work, we study the problem of personalized outfit recommendation, i.e.
automatically suggesting outfits to users that fit their personal fashion
preferences. Unlike existing recommendation systems that usually recommend
individual items, we suggest sets of items, which interact with each other, to
users. We propose a functional tensor factorization method to model the
interactions between user and fashion items. To effectively utilize the
multi-modal features of the fashion items, we use a gradient boosting based
method to learn nonlinear functions to map the feature vectors from the feature
space into some low dimensional latent space. The effectiveness of the proposed
algorithm is validated through extensive experiments on real world user data
from a popular fashion-focused social network.
[4]
Deep Multimodal Speaker Naming
Poster Session 2
/
Hu, Yongtao
/
Ren, Jimmy SJ.
/
Dai, Jingwen
/
Yuan, Chang
/
Xu, Li
/
Wang, Wenping
Proceedings of the 2015 ACM International Conference on Multimedia
2015-10-26
p.1107-1110
© Copyright 2015 ACM
Summary: Automatic speaker naming is the problem of localizing as well as identifying
each speaking character in a TV/movie/live show video. This is a challenging
problem mainly attributes to its multimodal nature, namely face cue alone is
insufficient to achieve good performance. Previous multimodal approaches to
this problem usually process the data of different modalities individually and
merge them using handcrafted heuristics. Such approaches work well for simple
scenes, but fail to achieve high performance for speakers with large appearance
variations. In this paper, we propose a novel convolutional neural networks
(CNN) based learning framework to automatically learn the fusion function of
both face and audio cues. We show that without using face tracking, facial
landmark localization or subtitle/transcript, our system with robust multimodal
feature extraction is able to achieve state-of-the-art speaker naming
performance evaluated on two diverse TV series. The dataset and implementation
of our algorithm are publicly available online.
[5]
Digital Rights Strategies in a Virtual World Marketplace
Electronic, Mobile and Ubiquitous Commerce
/
Hu, Yuanrong
/
Fan, Si
/
Wang, Qiuhong
HCIB 2015: 2nd International Conference on HCI in Business
2015-08-02
p.300-311
Keywords: Copyright strategy; Digital products; Virtual products
© Copyright 2015 Springer International Publishing Switzerland
Summary: This paper adopts the Heckman two-step model to analyze the impact of
copyright strategy on sales performance of digital product, using the panel
data comes from online virtual goods transaction website Xstreet.com. The
results show that (1) significant relationship exists between sales performance
and copyright strategy of digital product, but the influence of each copyright
strategy on sales performance are different; (2) A seller's copyright structure
within same product line can also affect sales performance of a digital
product, thus in order to optimize the copyright combinations, a seller should
fully consider the copyright strategy within the whole product line.
[6]
Energy and Performance of Smartphone Radio Bundling in Outdoor Environments
Technical Papers 2
/
Nika, Ana
/
Zhu, Yibo
/
Ding, Ning
/
Jindal, Abhilash
/
Hu, Y. Charlie
/
Zhou, Xia
/
Zhao, Ben Y.
/
Zheng, Haitao
Proceedings of the 2015 International Conference on the World Wide Web
2015-05-18
v.1
p.809-819
© Copyright 2015 ACM
Summary: Most of today's mobile devices come equipped with both cellular LTE and WiFi
wireless radios, making radio bundling (simultaneous data transfers over
multiple interfaces) both appealing and practical. Despite recent studies
documenting the benefits of radio bundling with MPTCP, many fundamental
questions remain about potential gains from radio bundling, or the relationship
between performance and energy consumption in these scenarios. In this study,
we seek to answer these questions using extensive measurements to empirically
characterize both energy and performance for radio bundling approaches. In
doing so, we quantify potential gains of bundling using MPTCP versus an ideal
protocol. We study the links between traffic partitioning and bundling
performance, and use a novel componentized energy model to quantify the energy
consumed by CPUs (and radios) during traffic management. Our results show that
MPTCP achieves only a fraction of the total performance gain possible, and that
its energy-agnostic design leads to considerable power consumption by the CPU.
We conclude that not only there is room for improved bundling performance, but
an energy-aware bundling protocol is likely to achieve a much better tradeoff
between performance and power consumption.
[7]
GalaxyExplorer: Influence-Driven Visual Exploration of Context-Specific
Social Media Interactions
Demonstrations
/
Liu, Xiaotong
/
Parthasarathy, Srinivasan
/
Shen, Han-Wei
/
Hu, Yifan
Companion Proceedings of the 2015 International Conference on the World Wide
Web
2015-05-18
v.2
p.215-218
© Copyright 2015 ACM
Summary: The ever-increasing size and complexity of social networks place a
fundamental challenge to visual exploration and analysis tasks. In this paper,
we present GalaxyExplorer, an influence-driven visual analysis system for
exploring users of various influence and analyzing how they influence others in
a social network. GalaxyExplorer reduces the size and complexity of a social
network by dynamically retrieving theme-based graphs, and analyzing users'
influence and passivity regarding specific themes and dynamics in response to
disaster events. In GalaxyExplorer, a galaxy-based visual metaphor is
introduced to simplify the visual complexity of a large graph with a
focus+context view. Various interactions are supported for visual exploration.
We present experimental results on real-world datasets that show the
effectiveness of GalaxyExplorer in theme-aware influence analysis.
[8]
Iterative Multi-View Hashing for Cross Media Indexing
Multimedia Search and Indexing
/
Hu, Yao
/
Jin, Zhongming
/
Ren, Hongyi
/
Cai, Deng
/
He, Xiaofei
Proceedings of the 2014 ACM International Conference on Multimedia
2014-11-03
p.527-536
© Copyright 2014 ACM
Summary: Cross media retrieval engines have gained massive popularity with rapid
development of the Internet. Users may perform queries in a corpus consisting
of audio, video, and textual information. To make such systems practically
possible for large mount of multimedia data, two critical issues must be
carefully considered: (a) reduce the storage as much as possible; (b) model the
relationship of the heterogeneous media data. Recently academic community have
proved that encoding the data into compact binary codes can drastically reduce
the storage and computational cost. However, it is still unclear how to
integrate multiple information sources properly into the binary code encoding
scheme.
In this paper, we study the cross media indexing problem by learning the
discriminative hashing functions to map the multi-view datum into a shared
hamming space. Not only meaningful within-view similarity is required to be
preserved, we also incorporate the between-view correlations into the encoding
scheme, where we map the similar points close together and push apart the
dissimilar ones. To this end, we propose a novel hashing algorithm called
Iterative Multi-View Hashing (IMVH) by taking these information into account
simultaneously. To solve this joint optimization problem efficiently, we
further develop an iterative scheme to deal with it by using a more flexible
quantization model. In particular, an optimal alignment is learned to maintain
the between-view similarity in the encoding scheme. And the binary codes are
obtained by directly solving a series of binary label assignment problems
without continuous relaxation to avoid the unnecessary quantization loss. In
this way, the proposed algorithm not only greatly improves the retrieval
accuracy but also performs strong robustness. An extensive set of experiments
clearly demonstrates the superior performance of the proposed method against
the state-of-the-art techniques on both multimodal and unimodal retrieval
tasks.
[9]
Matrix Completion for Cross-view Pairwise Constraint Propagation
Posters 1
/
Yang, Zheng
/
Hu, Yao
/
Liu, Haifeng
/
Chen, Huajun
/
Wu, Zhaohui
Proceedings of the 2014 ACM International Conference on Multimedia
2014-11-03
p.897-900
© Copyright 2014 ACM
Summary: As pairwise constraints are usually easier to access than label information,
pairwise constraint propagation attracts more and more attention in
semi-supervised learning. Most existing pairwise constraint propagation methods
are based on canonical graph propagation model, which heavily depends on the
edge weights in the graph and cannot preserve local and global consistency
simultaneously. In order to address this drawback, we cast cross-view pairwise
constraint propagation into a problem of low rank matrix completion and propose
a Matrix Completion method for cross-view Pairwise Constraint Propagation
(MCPCP). With low rank requirement and graph regularization, our MCPCP can
preserve local and global consistency simultaneously. We develop an algorithm
based on alternating direction method of multipliers (ADMM) to solve the
optimization problem. Finally, the effectiveness of MCPCP is demonstrated in
cross-view multimedia retrieval.
[10]
CONR: A Novel Method for Sentiment Word Identification
KM Track Posters
/
Liang, Jiguang
/
Zhou, Xiaofei
/
Hu, Yue
/
Guo, Li
/
Bai, Shuo
Proceedings of the 2014 ACM Conference on Information and Knowledge
Management
2014-11-03
p.1943-1946
© Copyright 2014 ACM
Summary: Sentiment word identification (SWI) is of high relevance to sentiment
analysis technologies and applications. Currently most SWI methods heavily rely
on sentiment seed words that have limited sentiment information. Even though
there emerge non-seed approaches based on sentiment labels of documents, but in
which the context information has not been fully considered. In this paper,
based on matrix factorization with co-occurrence neighbor regularization which
is derived from context, we propose a novel non-seed model called CONR for SWI.
Instead of seed words, CONR exploits two important factors: sentiment matching
and sentiment consistency for sentiment word identification. Experimental
results on four publicly available datasets show that CONR can outperform the
state of-the-art methods.
[11]
Evaluation of hands-on clinical exam performance using marker-less video
tracking
Health Care: HC13 -- Methodologies for Evaluating Health-Care Performance
/
Azari, David P.
/
Pugh, Carla M.
/
Laufer, Shlomi
/
Cohen, Elaine
/
Kwan, Calvin
/
Chen, Chia-Hsiung (Eric)
/
Yen, Thomas Y.
/
Hu, Yu Hen
/
Ray, Rebecca D.
/
Radwin, Robert G.
Proceedings of the Human Factors and Ergonomics Society 2014 Annual Meeting
2014-10-27
p.793-797
doi 10.1177/1541931214581145
© Copyright 2014 HFES
Summary: This study investigates the potential of using marker-less video tracking
for evaluating hands-on clinical skills. Experienced family practitioners
attending a national conference were recruited and asked to conduct a breast
examination on a simulator that presents different clinical pathologies. Videos
were taken of the clinician's hands during the exam. Video processing software
for tracking and quantifying hand motion kinematics was used. Videos were
divided into two segments: a general search segment and a mass exploration
segment. The general exploration segments exhibited motion patterns which
included 72% faster movement and 73% higher acceleration across clinical
pathologies. The most complex pathology exhibited 14% greater displacement for
pressing/rubbing than for general exploration. Marker-less video kinematic
tracking shows promise in discriminating between different examination
procedures, clinicians, and pathologies.
[12]
An Equation for Estimating Hand Activity Level Based on Measured Hand Speed
and Duty Cycle
Occupational Ergonomics (formerly Industrial Ergonomics): OE4 -- Ergonomics
Exposure Assessment
/
Akas, Oguz
/
Azari, David
/
Chen, Chia-Hsiung (Eric)
/
Hu, Yu Hen
/
Armstrong, Thomas J.
/
Ulin, Sheryl S.
/
Radwin, Robert G.
Proceedings of the Human Factors and Ergonomics Society 2014 Annual Meeting
2014-10-27
p.1600-1604
doi 10.1177/1541931214581333
© Copyright 2014 HFES
Summary: We are developing video processing algorithms for automatically measuring
the ACGIH TLV® hand activity level (HAL) using marker-less tracking of hand
movements. An equation for computing HAL ratings directly from tracked
kinematics, rather than using a frequency-duty cycle (DC) look-up table, more
readily lends itself to automated processing. Videos from the 33 Latko et al.
(1997) jobs were digitized and analyzed using marker-less tracking, and hand
root mean square (RMS) speed (S) was measured. A linear regression model was
developed for predicting the average observer rated HAL based on the measured
hand RMS speed and DC. Since the videos did not contain distance calibration,
speed was quantified in pixels/s and normalized by the distance of each
worker's hand breadth, measured in pixels. A Monte Carlo simulation was
performed using the US Army (1991) hand anthropometry data to determine how
variation is introduced in the equation as hand breadth varies. The resulting
equation was HAL= -1.06 + 0.0047 S + 0.053 DC and it predicted HAL ratings
within ±1. The development of an accurate equation for estimating HAL
ratings should enable use of automated and objective measurement in practice.
While expert observer HAL ratings offer speed and efficiency, use of objective
measurements based on worker hand kinematics should provide greater
reliability, as well as offering specific engineering aspects of the job that
may be addressed for reducing exposures and the risk of musculoskeletal
disorders. Furthermore, automated videos analysis may help improve the speed
and efficiency of making objective measurements in practice.
[13]
Developing early warning systems to predict students' online learning
performance
/
Hu, Ya-Han
/
Lo, Chia-Lun
/
Shih, Sheng-Pao
Computers in Human Behavior
2014-07
v.36
n.0
p.469-478
Keywords: Learning management system
Keywords: e-Learning
Keywords: Early warning system
Keywords: Data-mining
Keywords: Learning performance prediction
© Copyright 2014 Elsevier Ltd.
Summary: An early warning system can help to identify at-risk students, or predict
student learning performance by analyzing learning portfolios recorded in a
learning management system (LMS). Although previous studies have shown the
applicability of determining learner behaviors from an LMS, most investigated
datasets are not assembled from online learning courses or from whole learning
activities undertaken on courses that can be analyzed to evaluate students'
academic achievement. Previous studies generally focus on the construction of
predictors for learner performance evaluation after a course has ended, and
neglect the practical value of an "early warning" system to predict at-risk
students while a course is in progress. We collected the complete learning
activities of an online undergraduate course and applied data-mining techniques
to develop an early warning system. Our results showed that, time-dependent
variables extracted from LMS are critical factors for online learning. After
students have used an LMS for a period of time, our early warning system
effectively characterizes their current learning performance. Data-mining
techniques are useful in the construction of early warning systems; based on
our experimental results, classification and regression tree (CART),
supplemented by AdaBoost is the best classifier for the evaluation of learning
performance investigated by this study.
[14]
How to Improve Your Search Engine Ranking: Myths and Reality
/
Su, Ao-Jan
/
Hu, Y. Charlie
/
Kuzmanovic, Aleksandar
/
Koh, Cheng-Kok
ACM Transactions on The Web
2014-03
v.8
n.2
p.8
© Copyright 2014 ACM
Summary: Search engines have greatly influenced the way people access information on
the Internet, as such engines provide the preferred entry point to billions of
pages on the Web. Therefore, highly ranked Web pages generally have higher
visibility to people and pushing the ranking higher has become the top priority
for Web masters. As a matter of fact, Search Engine Optimization (SEO) has
became a sizeable business that attempts to improve their clients' ranking.
Still, the lack of ways to validate SEO's methods has created numerous myths
and fallacies associated with ranking algorithms.
In this article, we focus on two ranking algorithms, Google's and Bing's,
and design, implement, and evaluate a ranking system to systematically validate
assumptions others have made about these popular ranking algorithms. We
demonstrate that linear learning models, coupled with a recursive partitioning
ranking scheme, are capable of predicting ranking results with high accuracy.
As an example, we manage to correctly predict 7 out of the top 10 pages for 78%
of evaluated keywords. Moreover, for content-only ranking, our system can
correctly predict 9 or more pages out of the top 10 ones for 77% of search
terms. We show how our ranking system can be used to reveal the relative
importance of ranking features in a search engine's ranking function, provide
guidelines for SEOs and Web masters to optimize their Web pages, validate or
disprove new ranking features, and evaluate search engine ranking results for
possible ranking bias.
[15]
EDITED BOOK
Handbook of Human Centric Visualization
/
Huang, Weidong
2014
n.29
p.743
Springer New York
DOI: 10.1007/978-1-4614-7485-2
== Part I: Visual Communication ==
Visualizing Thought (3-40)
+ Tversky, Barbara
Gryphon: A 'Little' Domain-Specific Programming Language for Diffusion MRI Visualizations (41-61)
+ Chen, Jian
+ Cai, Haipeng
+ Auchus, Alexander P.
+ Laidlaw, David H.
Viewing Abstract Data as Maps (63-89)
+ Gansner, Emden R.
+ Hu, Yifan
+ Kobourov, Stephen G.
== Part II: Theory and Science ==
Individual Differences and Translational Science in the Design of Human-Centered Visualizations (93-113)
+ Green, Tera Marie
+ Arias-Hernandez, Richard
+ Fisher, Brian
Evaluating Visualization Environments: Cognitive, Social, and Cultural Perspectives (115-145)
+ Hundhausen, Christopher D.
On the Prospects for a Science of Visualization (147-175)
+ Rensink, Ronald A.
== Part III: Principles, Guidelines and Recommendations ==
Toward a Better Understanding and Application of the Principles of Visual Communication (179-201)
+ Bae, Juhee
+ Watson, Benjamin
Pep Up Your Time Machine: Recommendations for the Design of Information Visualizations of Time-Dependent Data (203-225)
+ Kriglstein, Simone
+ Pohl, Margit
+ Smuc, Michael
Using Textbook Illustrations to Extract Design Principles for Algorithm Visualizations (227-249)
+ Velázquez-Iturbide, J. Ángel
== Part IV: Methods ==
Conceptual Design for Sensemaking (253-283)
+ Blandford, Ann
+ Faisal, Sarah
+ Attfield, Simon
An Introduction and Guide to Evaluation of Visualization Techniques Through User Studies (285-313)
+ Forsell, Camilla
+ Cooper, Matthew
User-Centered Evaluation of Information Visualization Techniques: Making the HCI-InfoVis Connection Explicit (315-336)
+ Freitas, Carla M. D. S.
+ Pimenta, Marcelo S.
+ Scapin, Dominique L.
Eye Tracking on Visualizations: Progressive Extraction of Scanning Strategies (337-372)
+ Goldberg, Joseph H.
+ Helfman, Jonathan I.
Evaluating Overall Quality of Graph Visualizations Indirectly and Directly (373-390)
+ Huang, Weidong
Visual Analysis of Eye Tracking Data (391-409)
+ Raschke, Michael
+ Blascheck, Tanja
+ Burch, Michael
User Studies in Visualization: A Reflection on Methods (411-426)
+ Tory, Melanie
== Part V: Perception and Cognition ==
On the Benefits and Drawbacks of Radial Diagrams (429-451)
+ Burch, Michael
+ Weiskopf, Daniel
Measuring Memories for Objects and Their Locations in Immersive Virtual Environments: The Subjective Component of Memorial Experience (453-471)
+ Coxon, Matthew
+ Mania, Katerina
Human-Centric Chronographics: Making Historical Time Memorable (473-511)
+ Korallo, Liliya
+ Davis, Stephen Boyd
+ Foreman, Nigel
+ Moar, Magnus
Visualizing Multiple Levels and Dimensions of Social Network Properties (513-525)
+ McGrath, Cathleen
+ Blythe, Jim
+ Krackhardt, David
== Part VI: Dynamic Visualization ==
Adaptive Diagrams: A Research Agenda to Explore How Learners Can Manipulate Online Diagrams to Self-Manage Cognitive Load (529-550)
+ Agostinho, Shirley
+ Tindall-Ford, Sharon
+ Bokosmaty, Sahar
Dynamic Visualisations and Motor Skills (551-580)
+ Castro-Alonso, Juan Cristobal
+ Ayres, Paul
+ Paas, Fred
Dynamic Visualizations: A Two-Edged Sword? (581-604)
+ Lowe, Richard K.
Simultaneous and Sequential Presentation of Realistic and Schematic Instructional Dynamic Visualizations (605-622)
+ Nugteren, Michelle L.
+ Tabbers, Huib K.
+ Scheiter, Katharina
+ Paas, Fred
How Do You Connect Moving Dots? Insights from User Studies on Dynamic Network Visualizations (623-650)
+ Smuc, Michael
+ Federico, Paolo
+ Windhager, Florian
+ Aigner, Wolfgang
+ Zenk, Lukas
+ Miksch, Silvia
== Part VII: Interaction ==
Interaction Taxonomy for Tracking of User Actions in Visual Analytics Applications (653-670)
+ von Landesberger, Tatiana
+ Fiebig, Sebastian
+ Bremm, Sebastian
+ Kuijper, Arjan
+ Fellner, Dieter W.
Common Visualizations: Their Cognitive Utility (671-691)
+ Parsons, Paul
+ Sedig, Kamran
Distribution of Information Processing While Performing Complex Cognitive Activities with Visualization Tools (693-715)
+ Parsons, Paul
+ Sedig, Kamran
Human-Centered Interactivity of Visualization Tools: Micro- and Macro-level Considerations (717-743)
+ Sedig, Kamran
+ Parsons, Paul
+ Dittmer, Mark
+ Haworth, Robert
[16]
Efficient reconstruction, decomposition and editing for spatially-varying
reflectance data
Image-based rendering and tracking
/
Hu, Yong
/
Wang, Shan
/
Qi, Yue
Proceedings of the 2013 Conference on Graphics Interface
2013-05-29
p.55-62
© Copyright 2013 Authors
Summary: We present a new method for modeling real-world surface reflectance,
described with non-parametric spatially-varying bidirectional reflectance
distribution functions (SVBRDF). Our method seeks to achieve high
reconstruction accuracy, compactness and "editability" of representation
meanwhile speeding up the SVBRDF modeling processes. For a planar surface, we
1) design a capturing device to acquire reflectance samples at dense surface
locations; 2) propose a Laplacian-based angular interpolation scheme for a 2D
slice of BRDF at a given surface location, and then a Kernel Nyström
method for SVBRDF data matrix reconstruction; 3) propose a practical algorithm
to extract linear-independent basis BRDFs, and to calculate blending weights
through projecting reconstructed reflectance onto these bases. Results
demonstrate that our approach models real-world reflectance with both high
accuracy and high visual fidelity for real-time virtual environment rendering.
[17]
Whoo.ly: facilitating information seeking for hyperlocal communities using
social media
Papers: tensions in social media
/
Hu, Yuheng
/
Farnham, Shelly D.
/
Monroy-Hernández, Andrés
Proceedings of ACM CHI 2013 Conference on Human Factors in Computing Systems
2013-04-27
v.1
p.3481-3490
© Copyright 2013 ACM
Summary: Social media systems promise powerful opportunities for people to connect to
timely, relevant information at the hyper local level. Yet, finding the
meaningful signal in noisy social media streams can be quite daunting to users.
In this paper, we present and evaluate Whoo.ly, a web service that provides
neighborhood-specific information based on Twitter posts that were
automatically inferred to be hyperlocal. Whoo.ly automatically extracts and
summarizes hyperlocal information about events, topics, people, and places from
these Twitter posts. We provide an overview of our design goals with Whoo.ly
and describe the system including the user interface and our unique event
detection and summarization algorithms. We tested the usefulness of the system
as a tool for finding neighborhood information through a comprehensive user
study. The outcome demonstrated that most participants found Whoo.ly easier to
use than Twitter and they would prefer it as a tool for exploring their
neighborhoods.
[18]
A novel local patch framework for fixing supervised learning models
KM track: classification and semantic methods
/
Wang, Yilei
/
Wei, Bingzheng
/
Yan, Jun
/
Hu, Yang
/
Deng, Zhi-Hong
/
Chen, Zheng
Proceedings of the 2012 ACM Conference on Information and Knowledge
Management
2012-10-29
p.1233-1242
© Copyright 2012 ACM
Summary: In the past decades, machine learning models, especially supervised learning
algorithms, have been widely used in various real world applications. However,
no matter how strong a learning model is, it will suffer from the prediction
errors when it is applied to real world problems. Due to the black box nature
of supervised learning models, it is a challenging problem to fix the
supervised learning models by further learning from the failure cases it
generates. In this paper, we propose a novel Local Patch Framework (LPF) to
locally fix supervised learning models by learning from its predicted failure
cases. Since the learning models are generally globally optimized during
training process, our proposed LPF assumes that most of the learning errors are
led by local errors in the model. Thus we aim to break the black boxes of
learning models by identifying and fixing the local errors of various models
automatically. The proposed LPF has two key steps, which are local error region
subspace learning and local patch model learning. Through this way, we aim to
fix the errors of learning models locally and automatically with certain
generalization ability on unseen testing data. Experiments on both
classification and ranking problems show that the proposed LPF is effective and
outperforms the original algorithms and the incremental learning model.
[19]
Personalized document clustering with dual supervision
Search and sensemaking
/
Hu, Yeming
/
Milios, Evangelos E.
/
Blustein, James
/
Liu, Shali
Proceedings of the 2012 ACM Symposium on Document Engineering
2012-09-04
p.161-170
© Copyright 2012 ACM
Summary: The potential for semi-supervised techniques to produce personalized
clusters has not been explored. This is due to the fact that semi-supervised
clustering algorithms used to be evaluated using oracles based on underlying
class labels. Although using oracles allows clustering algorithms to be
evaluated quickly and without labor intensive labeling, it has the key
disadvantage that oracles always give the same answer for an assignment of a
document or a feature. However, different human users might give different
assignments of the same document and/or feature because of different but
equally valid points of view. In this paper, we conduct a user study in which
we ask participants (users) to group the same document collection into clusters
according to their own understanding, which are then used to evaluate
semi-supervised clustering algorithms for user personalization. Through our
user study, we observe that different users have their own personalized
organizations of the same collection and a user's organization changes over
time. Therefore, we propose that document clustering algorithms should be able
to incorporate user input and produce personalized clusters based on the user
input. We also confirm that semi-supervised algorithms with noisy user input
can still produce better organizations matching user's expectation
(personalization) than traditional unsupervised ones. Finally, we demonstrate
that labeling keywords for clusters at the same time as labeling documents can
improve clustering performance further compared to labeling only documents with
respect to user personalization.
[20]
Mining query subtopics from search log data
User intent
/
Hu, Yunhua
/
Qian, Yanan
/
Li, Hang
/
Jiang, Daxin
/
Pei, Jian
/
Zheng, Qinghua
Proceedings of the 35th Annual International ACM SIGIR Conference on
Research and Development in Information Retrieval
2012-08-12
p.305-314
© Copyright 2012 ACM
Summary: Most queries in web search are ambiguous and multifaceted. Identifying the
major senses and facets of queries from search log data, referred to as query
subtopic mining in this paper, is a very important issue in web search. Through
search log analysis, we show that there are two interesting phenomena of user
behavior that can be leveraged to identify query subtopics, referred to as 'one
subtopic per search' and 'subtopic clarification by keyword'. One subtopic per
search means that if a user clicks multiple URLs in one query, then the clicked
URLs tend to represent the same sense or facet. Subtopic clarification by
keyword means that users often add an additional keyword or keywords to expand
the query in order to clarify their search intent. Thus, the keywords tend to
be indicative of the sense or facet. We propose a clustering algorithm that can
effectively leverage the two phenomena to automatically mine the major
subtopics of queries, where each subtopic is represented by a cluster
containing a number of URLs and keywords. The mined subtopics of queries can be
used in multiple tasks in web search and we evaluate them in aspects of the
search result presentation such as clustering and re-ranking. We demonstrate
that our clustering algorithm can effectively mine query subtopics with an F1
measure in the range of 0.896-0.956. Our experimental results show that the use
of the subtopics mined by our approach can significantly improve the
state-of-the-art methods used for search result clustering. Experimental
results based on click data also show that the re-ranking of search result
based on our method can significantly improve the efficiency of users' ability
to find information.
[21]
Genre classification for million song dataset using confidence-based
classifiers combination
Poster abstracts
/
Hu, Yajie
/
Ogihara, Mitsunori
Proceedings of the 35th Annual International ACM SIGIR Conference on
Research and Development in Information Retrieval
2012-08-12
p.1083-1084
© Copyright 2012 ACM
Summary: We proposed a method to classify songs in the Million Song Dataset according
to song genre. Since songs have several data types, we trained sub-classifiers
by different types of data. These sub-classifiers are combined using both
classifier authority and classification confidence for a particular instance.
In the experiments, the combined classifier surpasses all of these
sub-classifiers and the SVM classifier using concatenated vectors from all data
types. Finally, the genre labels for the Million Song Dataset are provided.
[22]
Combining machine learning and human judgment in author disambiguation
Science, the past, and the future
/
Qian, Yanan
/
Hu, Yunhua
/
Cui, Jianling
/
Zheng, Qinghua
/
Nie, Zaiqing
Proceedings of the 2011 ACM Conference on Information and Knowledge
Management
2011-10-24
p.1241-1246
© Copyright 2011 ACM
Summary: Author disambiguation in digital libraries becomes increasingly difficult as
the number of publications and consequently the number of ambiguous author
names keep growing. The fully automatic author disambiguation approach could
not give satisfactory results due to the lack of signals in many cases.
Furthermore, human judgment on the basis of automatic algorithms is also not
suitable because the automatically disambiguated results are often mixed and
not understandable for humans. In this paper, we propose a Labeling Oriented
Author Disambiguation approach, called LOAD, to combine machine learning and
human judgment together in author disambiguation. LOAD exploits a framework
which consists of high precision clustering, high recall clustering, and top
dissimilar clusters selection and ranking. In the framework, supervised
learning algorithms are used to train the similarity functions between
publications and a clustering algorithm is further applied to generate
clusters. To validate the effectiveness and efficiency of the proposed LOAD
approach, comprehensive experiments are conducted. Comparing to conventional
author disambiguation algorithms, the LOAD yields much more accurate results to
assist human labeling. Further experiments show that the LOAD approach can save
labeling time dramatically.
[23]
Semantics-Enabled Policies for Information Sharing and Protection in the
Cloud
/
Hu, Yuh-Jong
/
Wu, Win-Nan
/
Yang, Jiun-Jan
Proceedings of the 2011 International Conference on Social Informatics
2011-10-06
p.198-211
Keywords: semantics-enabled policies; information sharing; data protection; national
security; cloud computing; privacy for social network cloud
© Copyright 2011 Springer
Summary: The cloud computing platform provides utility computing allowing people to
have convenient and flexible information sharing services on the web. We
investigate the inter-disciplinary area of information technology and law and
use semantics-enabled policies for modeling legal regulations in the cloud. The
semantics-enabled policies of information sharing and protection are
represented as a combination of ontologies and rules to capture the concept of
security and privacy laws. Ontologies are abstract knowledge representations of
information sharing and protection which extracted manually from the data
sharing and protection laws. Rules provide further enforcement power after
ontologies have been constructed. The emerging challenges of legalizing
semantics-enabled policies for laws in the cloud include mitigating the gap
between semantics-enabled policy and laws to avoid any ambiguity in the policy
representation, and resolving possible conflicts among policies when they are
required to integrate the laws from multiple jurisdictions.
[24]
Modeling spatially-varying reflectance based on Kernel Nyström
Poster session 1
/
Hu, Yong
/
Qi, Yue
/
Shen, Fangyang
Proceedings of the 2010 ACM Symposium on Virtual Reality Software and
Technology
2010-11-22
p.91-92
© Copyright 2010 ACM
Summary: We present a new method for modeling real-world surface reflectance,
described with non-parametric spatially-varying bidirectional reflectance
distribution functions (SVBRDF). Our method seeks to achieve high
reconstruction accuracy, compactness and "editability" of representation
meanwhile speeding up both the SVBRDF capturing and modeling processes. For a
planar surface, we 1) design a fast capturing device to acquire reflectance
samples at dense surface locations; 2) propose a Laplacian-based angular
interpolation scheme for a 2D slice of BRDF at a given surface location, and
then a Kernel Nyström method for SVBRDF data matrix reconstruction; 3)
propose a practical algorithm to extract linear-independent basis BRDFs, and to
calculate blending weights through projecting reconstructed reflectance onto
these bases. Results demonstrate that our approach models real-world
reflectance with both high accuracy and high visual fidelity for real-time
virtual environment rendering.
[25]
Emotion detection via discriminative kernel method
Cognitive systems and machine learning for assistive environments
/
Wang, Hua
/
Huang, Heng
/
Hu, Yanzi
/
Anderson, Mindi
/
Rollins, Pamela
/
Makedon, Fillia
Proceedings of the 3rd International Conference on PErvasive Technologies
Related to Assistive Environments
2010-06-23
p.7
Keywords: discriminative learning, emotion detection, facial expression recognition,
facial feature, kernel
© Copyright 2010 ACM
Summary: Human emotion detection is of substantial importance in diverse pervasive
applications in assistive environments. Because facial expressions provide a
key mechanism for understanding and conveying emotion, automatic emotion
detection through facial expression recognition has attracted increased
attention in both scientific research and practical applications in recent
years. Traditional facial expression recognition methods normally use only one
type of facial expression data, either static data extracted from one single
face image or motion dependent data obtained from dynamic face image sequences,
but seldom employ both. In this work, we propose a novel Discriminative Kernel
Facial Emotion Recognition (DKFER) method to integrate these two types of
facial expression data using a hybrid kernel, such that the advantages of both
of them are exploited. In addition, by using Linear Discriminant Analysis (LDA)
to transform the two types of original facial expression data into two more
discriminative lower-dimensional subspaces, the succeeding classification for
emotion detection can be carried out in a more efficient and effective way.
Encouraging experimental results in empirical studies demonstrate the practical
usage of the proposed DKFER method for emotion detection.
