Download presentation
Presentation is loading. Please wait.
1
16 February 2007
2
UMBC and Ebiquity UMBC is a research extensive University with a major focus on Information Technology Ebiquity is a large and active research group with the goal of “Building intelligent systems in open, heterogeneous, dynamic, distributed environments” Current research includes mobile and pervasive computing, security/trust/privacy, semantic web, multiagent systems, advanced databases, and high performance computing 5/1/2019
3
CSEE @ UMBC Computer Science & Electrical Engineering
UMBC’s largest department with 45 faculty, ~1000 undergrad, ~200 grad students Degree programs (graduate and undergraduate) Computer Science, Computer Engineering, Electrical Engineering Among research universities, UMBC is #1 in BS CS production and #18 in PhD production Many institutes, centers and labs Institute for Language and Information Technology, Center for Information Security and Assurance, Center for Photonics, Lab For Advanced Information Technology, VLSI Lab, CADIP, … Breadth and focus in research areas ~ $5M/year in sponsored research from Government and Industry Pervasive computing, AI, security, information retrieval, graphics, databases, VLSI, … 5/1/2019
4
5/1/2019
5
People and funding Faculty: Finin, Yesha, Joshi, Peng, Halem
Colleagues: Pinkston, Segall, … Students: ~10 PhD, ~10 MS, ~5 undergrad Funding: Current: DARPA (Trauma Pod, STTRs), NSF (two ITRs, Cybertrust, NSG, …), Intelligence community, NASA, NIST, Industry (IBM, Fujitsu, …) Recent: DARPA (CoABS, GENOA II, DAML), NSF (CAREER) 5/1/2019
6
Ebiquity Research Space
KR information extraction user modeling semantic web machine learning IR AI data mining Intelligent Information Systems web services/SOC DB knowledge management wearable computing mobility policies HPCC assurance Networking & Systems wireless Security trust context awareness DRM pervasive computing intrusion detection privacy 5/1/2019
7
Ebiquity Research Space
language technology robotics HCI planning KR Building intelligent systems in open, heterogeneous, dynamic, distributed environments user modeling semantic web data mining machine learning AI DB Intelligent Information Systems knowledge management web services IR service oriented computing wearable computing policies Networking & Systems wireless Security mobility assurance context awareness pervasive computing intrusion detection privacy trust 5/1/2019
8
Some Current Projects Managing spam in social media
Influence and opinions in social media Recognizing communities using opinionated links The Semantic Web Location based advertising Trustworthy data management in ad hoc networks 5/1/2019
9
Spam in the Blogosphere
Pranam Kolari Ph.D. Candidate
10
Spam in the Blogosphere
Types: comment spam, ping spam, spam blogs Akismet: “87% of all comments are spam” 75% of update pings are spam (ebiquity 2005) 20% of indexed blogs by popular blog search engines is spam (Umbria 2006, ebiquity 2006) “Spam blogs, sometimes referred to by the neologism splogs, are weblog sites which the author uses only for promoting affiliated websites” “Spings, or ping spam, are update pings sent from spam blogs” 1Wikipedia Filip Perich 5/1/2019
11
Plagiarized/Scraped Content
Profitable Contexts Plagiarized/Scraped Content Filip Perich 5/1/2019
12
Spam Pings – Inlets to Search Index Filip Perich 5/1/2019
13
Source of Splogs “Honestly, Do you think people who make $10k/month from adsense make blogs manually? Come on, they need to make them as fast as possible. Save Time = More Money! It's Common SENSE! How much money do you think you will save if you can increase your work pace by a hundred times? Think about it…” “Discover The Amazing Stealth Traffic Secrets Insiders Use To Drive Thousands Of Targeted Visitors To Any Site They Desire!” “Holy Grail Of Advertising... “ $ 197 “Easily Dominate Any Market, Any Search Engine, Any Keyword.” Filip Perich 5/1/2019
14
Splog Detection – Home Page Based
top features blogs splogs SVM based probabilistic splog detection (Kolari et al., 2006) Hand verified training set of blogs and splogs SVM Based Model Bag-of-words Binary Features Linear Kernel Precision/Recall of 87% Bag-of-words based feature using text on blog home-page Published at AAAI 2006 we what was my org flickr paper words weblog motion me thank go january trackback archives political find info news your another website best articles on perfect products uncategorized hot resources inc three copyright Filip Perich 5/1/2019
15
56% of all blogs are splogs! Silicon Valley and Splog Valley?
1. Mountain View, CA 2. Washington DC 3. San Francisco, CA 4. Orlando, FL 5. Lansing, MI High PPC contexts are primary spam drivers auto buy california cancer card casino cheap consolidation credit debt diet discount equipment estate finance florida forex free gift golf health hotel insurance jewelry lawyer loan loans medical money mortgage new online phone poker rental sale software texas trading travel used vacation video wedding Filip Perich 5/1/2019
16
Influence and Opinions in Social Media
Akshay Java Ph.D. Candidate
17
Biz Intelligence from Social Media
“Social media describes the online tools and platforms that people use to share opinions, insights, experiences, and perspectives with each other.” – Wikipedia, Feb 07 Social Media is a dynamic and growing area, that includes blogs, wikis, forums, photo and video sharing sites, etc. Goal: Building scalable, solutions and frameworks for analyzing high volumes of blog data to derive business intelligence Filip Perich 5/1/2019
18
Knowing & Influencing your Market
Your goal is to market Apple’s iPod phone How can you track the buzz about it? What are the relevant communities and blogs? Which communities are fans, which are suspicious, which are put off by the hype? Is your advertising having an effect? The desired effect? Which bloggers are influential in this market? Of these, which are already onboard and which are lost causes? To whom should you send details or evaluation samples? Filip Perich 5/1/2019
19
Influence Detection Influence Detection Influence, bias in MSM
Often buyers look for opinions and reviews on blogs Detecting influential nodes and their role in how people perceive a service is an important tool for marketing Using topic, social structure, opinions, biases and temporal information we can develop an accurate model for influence Influence, bias in MSM Top Democrat MSM Sources Top Republican MSM Sources Filip Perich 5/1/2019
20
Opinions in Social Media
TREC 06: Finding opinionated posts, either positive or negative, about a query 2006 TREC Blog corpus: 80K blogs 300K posts 50 test queries Challenges: open domain sentiment words, slangs, subject Reader’s Perspective “Starbucks Sandwiches are bad!” “I went to school early so I would have time to grab some lunch. Which ended up consisting of a crappy sandwich from starbucks and a chai latte. Lacey came into Starbucks while I was there so we chatted for a little bit and she thought that I might be in her class. After I finished eating I headed to school and checked the board……..”1 Narrative Expressed Opinions Opinions can effect buying decisions of customers [1] Filip Perich 5/1/2019
21
Finding Feeds That Matter
Analysis of Bloglines Feeds 83K publicly listed subscribers 2.8M feeds, 500K are unique 26K users (35%) use folders to organize subscriptions Data collected in May 2006 Before Merge Top Advertising Feeds 1. Adrants » Marketing and Advertising News With Attitude 2. Adverblog: advertising and new media marketing 3. 4. adfreak 5. AdJab 6. MIT Advertising Lab: future of advertising and advertising technology 7. AdPulp: Daily Juice from the Ad Biz 8. Advertising/Design Goodness Related Tags: advertising marketing media news design After Merge Filip Perich 5/1/2019
22
Anubhav Kale M.S Student
Link Polarity Anubhav Kale M.S Student
23
Modeling Influence Using Link Polarity
Motivation Growing interest in exploring role of communities in social media Community detection algorithms rely more on link structure and less on sentiment associated with links Convert a sparsely connected blog graph into a densely connected graph with sentiment weight attached to every link Approach Link Polarity : Analyze text surrounding html links in blog posts to determine bias of bloggers about each other Trust Propagation: Use trust propagation models to spread the polarity from a small subset of “connected” bloggers to all bloggers. Experiments Political blogosphere as a specific domain of study. Goal is to divide the dataset of political blogs into left and right leaning blogs. Detection of leaning based on positive/neural/negative score from influential bloggers (high in-link blogs) in both communities Validation with a hand-labeled dataset indicates around 60% correct classification Filip Perich 5/1/2019
24
Anubhav Kale M.S Student
Semantic Web Anubhav Kale M.S Student
25
http://swoogle.umbc.edu/ Running since summer 2004
1.8M RDF docs, 320M triples, 10K ontologies, 15K namespaces, 1.3M classes, 175K properties, 43M instances, 600 registered users Filip Perich 5/1/2019 27
26
Applications and Use Cases
1 Supporting Semantic Web Developers Ontology designers, vocabulary discovery, who’s using my ontologies or data?, use analysis, errors, statistics, etc. Searching Specialized Collections Spire: aggregating observations and data from biologists InferenceWeb: searching over and enhancing proofs Supporting Semantic Web Tools Triple shop: finding data for SPARQL queries 2 3 Filip Perich 5/1/2019 28
27
What are body masses of fishes that eat fishes?
Swoogle Triple Shop What are body masses of fishes that eat fishes? At UMBC we have designed a TripleShop, a workshop for semantic web data. We write a database query in a language called SPARQL Essentially, we are writing patterns that will match the kinds of triples I just showed you. Roughly, what we’re saying here is: Find actual foodweb links and consider the taxa that are predators and the taxa that are prey, and make sure that these predators and prey are both fish, and give me the maximum body sizes of these prey and their predators. IF you are familiar with relational database queries, you might notice that I haven’t specified what data tables the data should come from. . . . leaving out the FROM clause Filip Perich 5/1/2019 29
28
Location Based Advertising
Olga Ratsimor Ph.D. Candidate
29
Intelligent Marketing in Mobile Peer-To-Peer Environments eNcentive
MH2 MH1 MH3 Animated Slide Filip Perich 5/1/2019
30
eNcentive: Targeting Locations
Filip Perich 5/1/2019
31
eNcentive + TrueBahn Trust based mobile shopping
Mobile users are divided into two groups Kids and Adults Groups are interlined Kids have a circle of friends that can grow and shrink Adults have circle of other Adults that they trust. The circle can grow and shrink. Kids can shop for items that thy have been authorized to buy If there is a new item on the wish list then kids need to find an Adult from their trust network to give them authorization. Filip Perich 5/1/2019
32
Bartering of Digital Goods and Services in Pervasive Environments
Context-based, continuous bartering with peers in pervasive environments Value in Use - is the value of the particular electronic good or service for the particular user. Value in Exchange - reflects the potential value of the service against any other service in value-for-value exchanges. Effects of collaborative strategies on the welfare of the network Effects of digital rights management restrictions on bartering collaborations Exploiting relationships between groups of devices Filip Perich 5/1/2019
33
Jesse English and Sandor Dornbush
With help from: Dr. Tim Oates, Dr. Anupam Joshi, Dr. Zary Segall & Chad Eby. Filip Perich 5/1/2019
34
Activity Aware Music Feedback loop between the user and their music.
Use the users physiological state to influence the music selection. The music influences the users mood and mindset. Filip Perich 5/1/2019
35
Current XPod Platform Nokia 5500 Sport Phone Embedded Accelerometers
MP3 Player Filip Perich 5/1/2019
36
Current Projects Incorporate song metadata
Human Generated, eg Last.fm Machine Generated Incorporate other meta information Location Information Recent Phone Calls Weather Filip Perich 5/1/2019
37
Advertising Potential
“Cellphone Ads May Take Off Soon,” New York Times Febuary 14th, 2007 Target adds based on activity profiles. Send running adds to somebody who runs. Target adds based on musical tastes. Truck adds for somebody who listens to country. Filip Perich 5/1/2019
38
Trustworthy Data Management in Ad Hoc Networks
Anand Patwardhan Ph.D Candidate Jim Parker Ph.D Candidate
39
Situation Awareness allows Adaptation
GPS satellite Localized and distributed Wireless Access points Hazard warnings, Detours, Inclement weather, Road conditions, Traveler info. Localized Info-Stream Services Various forms of connectivity Location & directions GSM, GPRS, EDGE, E-VDO WiMax GPS VANET connectivity Update propagation Onboard Computer with various sensors: GPS location Cameras Engine Condition Tire pressure etc. Situation Awareness allows Adaptation Filip Perich 5/1/2019
40
Securing MANETs Security for resources Trust in other resources
Malicious behavior (Activity monitoring) Misuse (Resource protection) Response/recourse (Accountability) Trust in other resources Dependence on recommendations (Identities and Reputations) Reliability of information Filip Perich 5/1/2019
41
Trust evolution, reputation management,
Cross-layer Analysis Trust evolution, reputation management, recourse Intrusion Detection Application Commendations Accusations (to other devices) Packet dropping, Mangling, injection Transport Routing attacks, disruptions Link Unfair contention, Jamming MAC/PHY Response Filip Perich 5/1/2019
42
Simulation and Modeling
Identity Cryptographic Addresses Mobility, congestion, radio interference Reduce false positives Scalability Large radio-ranges or dense networks Aggregation of data Communicate intrusions data to warn others Filip Perich 5/1/2019
43
Other Current and Recent Projects
44
Other Current and Recent Projects
Pervasive and mobile computing (1) Trauma Pod (2) Context aware pervasive computing (3) Mogatu: Tivo for mobile computing Semantic Web (4) Swoogle: searching and indexing Semantic Web data (5) Semnews: text understanding and extraction (6) Agents and the Semantic Web (7) Spire: Semantic Web for data discovery and integration Security and trust (8) Semantic policy languages (9) Semdis: Discovering Semantic Links (10) Securing ad hoc networks (11) Privacy for passive RFID tags Information extraction and retrieval (12) Recognizing spam weblogs (13) Extracting opinions from weblogs (14) Modeling the Spread of Influence on the Blogosphere 5/1/2019
45
5/1/2019
46
5/1/2019
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.