SLIDE 1
Date: 2017/05/09 Author: Mayank Kejriwal, Pedro Szekely Source: ACM WWW’ 17 Advisor: Jia-ling Koh Speaker : Yi-hui Lee
1
Outline
- Introduction
- Approach
- Experiment
- Conclusion
2
Information Extraction in Illicit Web Domains Date: 2017/05/09 - - PowerPoint PPT Presentation
Information Extraction in Illicit Web Domains Date: 2017/05/09 - - PowerPoint PPT Presentation
Information Extraction in Illicit Web Domains Date: 2017/05/09 Author: Mayank Kejriwal, Pedro Szekely Source: ACM WWW 17 Advisor: Jia-ling Koh Speaker : Yi-hui Lee 1 Outline Introduction Approach Experiment Conclusion 2
SLIDE 2
SLIDE 3
Introduction
- Information Extraction:
3
SLIDE 4
Introduction(cont.)
- Information Extraction on Dark web(human trafficking):
4
ages (of human trafficking victims) locations prices of services posting dates
SLIDE 5
- A high-level overview of the proposed information
extraction approach:
5
preprocessing
step 1 step 2
Apply recognizers
step 3
word representation learning supervised classifier
step 4
- utput:Annotated corpus
input: Dark Web
Introduction(cont.)
SLIDE 6
Outline
- Introduction
- Approach
- Experiment
- Conclusion
6
SLIDE 7
Approach
- Step 1. Preprocessing:
Readability Text Extractor(RTE):
- > Mercury Web Parser
NLTK:
- RTE string output -> sentence tokenize -> word tokenize
- > list of tokens
7
SLIDE 8
Approach(cont.)
8
- Step 2. Apply recognizers:
GeoNames-Cities GeoNames-States RegEx-Ages: use regular expressions Dictionary-Names: person names
SLIDE 9
Approach(cont.)
9
- Step 3. Word Representation learning:
D1: The cow is in the farm. D2: I jumped over the farm. D3: I saw a cow in the farm.
D1 D2 D3 The 1 cow 1 1 jumped 1
- ver
1 the 1 1 1 moon farm 1 1 1
sim(cow, farm) = 2/(sqrt(2)+sqrt(3)) = 0.64 sim(cow, moon) = 0
SLIDE 10
Approach(cont.)
10
- Step 3. Word Representation learning:
Random Index [27]
- randomly assigned -1, 0, 1 to the vector’s attribute
SLIDE 11
Approach(cont.)
11
- Step 4. Supervised Contextual Classifier:
Aggregate vectors -> l2-normalization
I saw a cow jumped over the farm
saw = [1, 0, 0, …, 1, 0] a = [1, 1, 1, …, 1, 1] cow = [1, 0, 1, …, 0, 0] jumped = [0, 0, 0, …, 1, 1]
- ver = [0, 0, 1, …, 0, 1]
aggregate = [1, 0, 0, …, 1, 0, 1, 1, 1, …, 1, 1, ……, 0, 1] l2-normalization = [0.0001, 0, 0, …, 0.0001, 0, 0.0001, 0.0001, 0.0001, …, 0.0001, 0.0001, ……, 0.0000, 1]
SLIDE 12
Approach(cont.)
12
- Step 4. Supervised Contextual Classifier:
Classifier: Random forest
SLIDE 13
Outline
- Introduction
- Approach
- Experiment
- Conclusion
13
SLIDE 14
Experiment
14
- Datasets and Ground-truths:
Research conducted in the DARPA MEMEX program
- Ground-truths:
SLIDE 15
Experiment(cont.)
- Baselines:
Stanford Named Entity Recognition system (NER)
15
SLIDE 16
- Evaluation:
Experiment(cont.)
16
SLIDE 17
Experiment(cont.)
- Feature selection:
17
SLIDE 18
Outline
- Introduction
- Approach
- Experiment
- Conclusion
18
SLIDE 19
Conclusion
19
- We presented a lightweight, feature-agnostic
Information Extraction approach that is suitable for illicit Web domains.
- Our approach relies on unsupervised derivation of
word representations from an initial corpus, and the training of a supervised contextual classifier using external high-recall recognizers and a handful of manually verified annotations.
- Real-world settings:
End Human Trafficking hackathon organized by the office of the District Attorney of New York17