SlideShare una empresa de Scribd logo
1 de 36
NEGOBOT
A conversational agent based on game
theory for the detection of paedophile
behaviour
Children have become active users of the Internet
One of the worst problems in cyber-society is
Commercial systems
analyse conversations
to automatically classify them
Our approach?
Meet:
NEGOBOT
Objective:
To detect paedophile
behaviour
As a chatter bot, negobot “knows” about:
Natural Language Processing
Information Retrieval
Automatic Learning
Game theory
Negobot’s
architecture
AI’s system knowledge
Gathering groups of
representative conversations
considered offensive.
http://www.perverted-justice.com
377 conversations
We use Lucene in order
to rank how similar are
Negobot’s conversations
with actual paedophile’s
conversations
A system to
understand the conversations
Replacing
“emoticons”
SMS-like wording
translation
Correcting
misspelled words
Question-answering
patterns
(AIML)
Random response
waiting times
Colloquial and
SMS-like language
Forced
language errors
Game theory
A structure of seven
chatterbots, with different
behaviours
Conversation level
An evaluation function
to classify, in real time, the
current conversation
Functional
flow
EXAMPLES
Passive
conversation
Aggresive
conversation
Limitations?
The key is the
language
Future?
WSD, opinion
mining, …
Improve
AIML
Collaborative
agents
Working with the Spanish’
Cyber-crime unit…
…trying to find those
monsters
References
1. Little girl: http://4.bp.blogspot.com/-qoMi9XA-
pfE/UD4Il8NOF3I/AAAAAAAADH4/Dy83sETvTgI/s0/Bank+Interview+Tips.jpg
2. Predator: http://1.bp.blogspot.com/-
ZkA7FRuhLu8/TouRfRHOwmI/AAAAAAAAahc/9auIEO8M1m4/s400/pedofilia%2B9%
255B5%255D.jpg
3. Conversation icon: http://www.vendorregistry.com/images/home-
slides/conversation-icon.png?sfvrsn=0
4. Lighthouse: http://lucaskrech.com/blog/wp-
content/uploads/2010/04/lighthouse4tracing.jpg
5. Human brain: http://www.whyworrybook.com/wp-
content/uploads/2013/01/canstockphoto1694623-2-brain-with-shooting-lines.jpg
6. Reveal-listen-understanding: http://2.bp.blogspot.com/-HXGhx9-
CNts/UhWpYnBvSDI/AAAAAAAAAIE/Ic7EPi-f94A/s1600/understanding.jpg
7. Chess: http://2.bp.blogspot.com/-
5_3295FDOd4/UcQvY6s05uI/AAAAAAAAa4I/IF9Pf_Qxa2w/s1600/Chess+HD+Picture
s7.jpg
8. Prison: http://www.ereverev.co.il/UploadImg/Articles/12826.jpg

Más contenido relacionado

La actualidad más candente

Semantic webslideshareversion
Semantic webslideshareversionSemantic webslideshareversion
Semantic webslideshareversion
Caroline_Rose
 

La actualidad más candente (12)

Ocr algorithm for ge’ez characters
Ocr algorithm for ge’ez charactersOcr algorithm for ge’ez characters
Ocr algorithm for ge’ez characters
 
Cognitive services and intelligent chatbots
Cognitive services and intelligent chatbotsCognitive services and intelligent chatbots
Cognitive services and intelligent chatbots
 
AI Overview
AI OverviewAI Overview
AI Overview
 
Artificial intelligence
Artificial intelligenceArtificial intelligence
Artificial intelligence
 
Artificial Intelligence for Speech Recognition
Artificial Intelligence for Speech RecognitionArtificial Intelligence for Speech Recognition
Artificial Intelligence for Speech Recognition
 
ARTIFICIAL INTELLIGENCE
ARTIFICIAL INTELLIGENCEARTIFICIAL INTELLIGENCE
ARTIFICIAL INTELLIGENCE
 
Group 2 -innovation in smartphones-
Group 2 -innovation in smartphones-Group 2 -innovation in smartphones-
Group 2 -innovation in smartphones-
 
Semantic webslideshareversion
Semantic webslideshareversionSemantic webslideshareversion
Semantic webslideshareversion
 
Speech recognition an overview
Speech recognition   an overviewSpeech recognition   an overview
Speech recognition an overview
 
AI programming languages
AI programming languagesAI programming languages
AI programming languages
 
Noa Ha'aman - 2017 - MojiSem: Varying Linguistic Purposes of Emoji in (Twitte...
Noa Ha'aman - 2017 - MojiSem: Varying Linguistic Purposes of Emoji in (Twitte...Noa Ha'aman - 2017 - MojiSem: Varying Linguistic Purposes of Emoji in (Twitte...
Noa Ha'aman - 2017 - MojiSem: Varying Linguistic Purposes of Emoji in (Twitte...
 
Chatbots and AI
Chatbots and AIChatbots and AI
Chatbots and AI
 

Destacado (6)

SXSW: Brands That Believe in Sex After Marriage
SXSW: Brands That Believe in Sex After MarriageSXSW: Brands That Believe in Sex After Marriage
SXSW: Brands That Believe in Sex After Marriage
 
Game theory and 3G License Auction
Game theory and 3G License AuctionGame theory and 3G License Auction
Game theory and 3G License Auction
 
Change Detection Dubai
Change Detection DubaiChange Detection Dubai
Change Detection Dubai
 
Game theory
Game theoryGame theory
Game theory
 
PCA and Classification
PCA and ClassificationPCA and Classification
PCA and Classification
 
Application Of Game Theory
Application Of Game TheoryApplication Of Game Theory
Application Of Game Theory
 

Similar a Negobot: A conversational agent based on game theory for the detection of paedophile behaviour - CISIS 2012

Artificial Intelligence
Artificial IntelligenceArtificial Intelligence
Artificial Intelligence
iarthur
 
New Approaches at Natural Language Processing Systems
New Approaches at Natural Language Processing SystemsNew Approaches at Natural Language Processing Systems
New Approaches at Natural Language Processing Systems
Andrejkovics Zoltán
 
418 01 context for ai
418 01 context for ai418 01 context for ai
418 01 context for ai
studycs
 

Similar a Negobot: A conversational agent based on game theory for the detection of paedophile behaviour - CISIS 2012 (20)

Laorden 2012 cisis_negobot
Laorden 2012 cisis_negobotLaorden 2012 cisis_negobot
Laorden 2012 cisis_negobot
 
How to make a chatbot more human
How to make a chatbot more humanHow to make a chatbot more human
How to make a chatbot more human
 
ms_3.pdf
ms_3.pdfms_3.pdf
ms_3.pdf
 
Sumit A
Sumit ASumit A
Sumit A
 
Artificial Intelligence
Artificial IntelligenceArtificial Intelligence
Artificial Intelligence
 
Reporting Metasystem Design and Penalization Strategy Best Practices (Present...
Reporting Metasystem Design and Penalization Strategy Best Practices (Present...Reporting Metasystem Design and Penalization Strategy Best Practices (Present...
Reporting Metasystem Design and Penalization Strategy Best Practices (Present...
 
DeepPavlov 2019
DeepPavlov 2019DeepPavlov 2019
DeepPavlov 2019
 
Microsoft AI at SAI '17
Microsoft AI at SAI '17Microsoft AI at SAI '17
Microsoft AI at SAI '17
 
Can AI compete with a smile? nicola strong srai presentation 14 september 2016
Can AI compete with a smile? nicola strong srai presentation   14 september 2016Can AI compete with a smile? nicola strong srai presentation   14 september 2016
Can AI compete with a smile? nicola strong srai presentation 14 september 2016
 
Artificial intelligence
Artificial intelligenceArtificial intelligence
Artificial intelligence
 
PPT Emotion Sensor.pptx
PPT Emotion Sensor.pptxPPT Emotion Sensor.pptx
PPT Emotion Sensor.pptx
 
From ELIZA to Alexa and Beyond
From ELIZA to Alexa and BeyondFrom ELIZA to Alexa and Beyond
From ELIZA to Alexa and Beyond
 
How to choose Chatbot platform & wtf is ML, DL, NLP, NLU, CA
How to choose Chatbot platform & wtf is ML, DL, NLP, NLU, CAHow to choose Chatbot platform & wtf is ML, DL, NLP, NLU, CA
How to choose Chatbot platform & wtf is ML, DL, NLP, NLU, CA
 
Ml in games intel game developer presentation v1.2
Ml in games intel game developer presentation v1.2Ml in games intel game developer presentation v1.2
Ml in games intel game developer presentation v1.2
 
Chatbot
ChatbotChatbot
Chatbot
 
Natural Language Processing Solution
Natural Language Processing SolutionNatural Language Processing Solution
Natural Language Processing Solution
 
Looking at Machine Learning in Games
Looking at Machine Learning in GamesLooking at Machine Learning in Games
Looking at Machine Learning in Games
 
New Approaches at Natural Language Processing Systems
New Approaches at Natural Language Processing SystemsNew Approaches at Natural Language Processing Systems
New Approaches at Natural Language Processing Systems
 
Beekman5 std ppt_14
Beekman5 std ppt_14Beekman5 std ppt_14
Beekman5 std ppt_14
 
418 01 context for ai
418 01 context for ai418 01 context for ai
418 01 context for ai
 

Más de Carlos Laorden

Más de Carlos Laorden (10)

Collective Classification for Packed Executable Identification - CEAS 2011
Collective Classification for Packed Executable Identification - CEAS 2011Collective Classification for Packed Executable Identification - CEAS 2011
Collective Classification for Packed Executable Identification - CEAS 2011
 
Collective classification for unknown malware detection - SECRYPT 2011
Collective classification for unknown malware detection - SECRYPT 2011Collective classification for unknown malware detection - SECRYPT 2011
Collective classification for unknown malware detection - SECRYPT 2011
 
Puma: Permission Usage to detect Malware in Android - CISIS 2012
Puma: Permission Usage to detect Malware in Android - CISIS 2012Puma: Permission Usage to detect Malware in Android - CISIS 2012
Puma: Permission Usage to detect Malware in Android - CISIS 2012
 
A Threat Model Approach to Threats and Vulnerabilities in On-line Social Netw...
A Threat Model Approach to Threats and Vulnerabilities in On-line Social Netw...A Threat Model Approach to Threats and Vulnerabilities in On-line Social Netw...
A Threat Model Approach to Threats and Vulnerabilities in On-line Social Netw...
 
Anomaly Detection using String Analysis for Android Malware Detection - CISIS...
Anomaly Detection using String Analysis for Android Malware Detection - CISIS...Anomaly Detection using String Analysis for Android Malware Detection - CISIS...
Anomaly Detection using String Analysis for Android Malware Detection - CISIS...
 
Twitter Content-based Spam Filtering - CISIS 2013
Twitter Content-based Spam Filtering - CISIS 2013Twitter Content-based Spam Filtering - CISIS 2013
Twitter Content-based Spam Filtering - CISIS 2013
 
Enhancing Scalability in Anomaly-based Email Spam Filtering - CEAS 2011
Enhancing Scalability in Anomaly-based Email Spam Filtering - CEAS 2011Enhancing Scalability in Anomaly-based Email Spam Filtering - CEAS 2011
Enhancing Scalability in Anomaly-based Email Spam Filtering - CEAS 2011
 
Collective Classification for Spam Filtering - CISIS 2011
Collective Classification for Spam Filtering - CISIS 2011Collective Classification for Spam Filtering - CISIS 2011
Collective Classification for Spam Filtering - CISIS 2011
 
On the Study of Anomaly-based Spam Filtering Using Spam as Representation of ...
On the Study of Anomaly-based Spam Filtering Using Spam as Representation of ...On the Study of Anomaly-based Spam Filtering Using Spam as Representation of ...
On the Study of Anomaly-based Spam Filtering Using Spam as Representation of ...
 
Anomaly-based Spam Filtering - SECRYPT 2011
Anomaly-based Spam Filtering - SECRYPT 2011Anomaly-based Spam Filtering - SECRYPT 2011
Anomaly-based Spam Filtering - SECRYPT 2011
 

Último

+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
?#DUbAI#??##{{(☎️+971_581248768%)**%*]'#abortion pills for sale in dubai@
 

Último (20)

Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobe
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, AdobeApidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobe
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobe
 
Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...
 
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherStrategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a Fresher
 
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
 
Real Time Object Detection Using Open CV
Real Time Object Detection Using Open CVReal Time Object Detection Using Open CV
Real Time Object Detection Using Open CV
 
TrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
TrustArc Webinar - Unlock the Power of AI-Driven Data DiscoveryTrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
TrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
 
DBX First Quarter 2024 Investor Presentation
DBX First Quarter 2024 Investor PresentationDBX First Quarter 2024 Investor Presentation
DBX First Quarter 2024 Investor Presentation
 
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
 
Apidays Singapore 2024 - Modernizing Securities Finance by Madhu Subbu
Apidays Singapore 2024 - Modernizing Securities Finance by Madhu SubbuApidays Singapore 2024 - Modernizing Securities Finance by Madhu Subbu
Apidays Singapore 2024 - Modernizing Securities Finance by Madhu Subbu
 
AXA XL - Insurer Innovation Award Americas 2024
AXA XL - Insurer Innovation Award Americas 2024AXA XL - Insurer Innovation Award Americas 2024
AXA XL - Insurer Innovation Award Americas 2024
 
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
 
GenAI Risks & Security Meetup 01052024.pdf
GenAI Risks & Security Meetup 01052024.pdfGenAI Risks & Security Meetup 01052024.pdf
GenAI Risks & Security Meetup 01052024.pdf
 
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024
 
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
 
Ransomware_Q4_2023. The report. [EN].pdf
Ransomware_Q4_2023. The report. [EN].pdfRansomware_Q4_2023. The report. [EN].pdf
Ransomware_Q4_2023. The report. [EN].pdf
 
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
Apidays New York 2024 - The Good, the Bad and the Governed by David O'Neill, ...
 
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
 
MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024MINDCTI Revenue Release Quarter One 2024
MINDCTI Revenue Release Quarter One 2024
 
presentation ICT roal in 21st century education
presentation ICT roal in 21st century educationpresentation ICT roal in 21st century education
presentation ICT roal in 21st century education
 
Automating Google Workspace (GWS) & more with Apps Script
Automating Google Workspace (GWS) & more with Apps ScriptAutomating Google Workspace (GWS) & more with Apps Script
Automating Google Workspace (GWS) & more with Apps Script
 

Negobot: A conversational agent based on game theory for the detection of paedophile behaviour - CISIS 2012

Notas del editor

  1. A chatter bot that poses as a kid in chats, social networks and similar services on the Internet to detect paedophile behaviour.
  2. Negobot includes the use of different NLP techniques, chatter-bot technologies and game theory for the strategical decision making. Finally, the glue that binds them all is an evaluation function, which in fact determines how the child emulated by the conversational agent behaves.
  3. Firstwehadtogatherrepresentativeconversations, consideredoffensive.That knowledge came from the website Perverted Justice.
  4. This website offers an extensive database of paedophile conversations with victims, used in other research works.A total of 377 real conversations were chosen to populate our database.Besides, Perverted Justice users provide an evaluation of each conversation's seriousness by selecting a level of “slimyness”, that is, how disgusting the conversation is. Note that this evaluation is given by the website's visitors, so it may not be accurate, but we consider that it is a proper baseline in order to compare future conversations of the chatter-bot.
  5. We use Lucene, a high-performance Information Retrieval tool, to stablish how similar are Negobot’s conversations with those conversations retrieved from perverted justice.
  6. to hide the real nature of chatterbot.This system can translate the words from this SMS language to normal and correct language and viceversa.
  7. The system replaces “emoticons” and misspelled words are corrected through Levenshtein distance
  8. Negobot uses the Artifiial Intelligence Markup Language (AIML) to provide the bot with the capacity of giving consistent answers and, also, the ability to be an active part in the conversation and to start new topics or discussions about the subject's answers.Although the AIML structure is based on the Galaia project, which has successfully implanted derived projects in social networks and chat systems [4, 5, 7], we edited their AIML les to adequate them to our needs. Those les can be found at the authors' website
  9. An identification and fitness system inside the conversations able to maintain a normal conversation flow like a correct conversation between two real persons.
  10. to hide the real nature of chatterbot.This system can translate the words from this SMS language to normal and correct language and viceversa.
  11. *Initial state (Start level or Level 0). In this level, the conversation has started recently or it is within the fixed limits. The user can stay indefinitely in this level if the conversation does not contain disturbing content. The topics of conversation are trivial and the provided information about the bot is brief: only the name, age, gender and home-town. The bot does not provide more personal information until higher levels.*Possibly not (Level -1). In this level, the subject talking to the bot, does not want to continue the conversation. Since this is the first negative level, the bot will try to reactivate the conversation. To this end, the bot will ask for help about family issues, bullying or other types of adolescent problems.*Probably not (Level -2). In this level, the user is too tired about the conversation and his language and ways to leave it are less polite than before. The conversation is almost lost. The strategy in this stage is to act as a victim to which nobody pays any attention, looking for affection from somebody.*Is not a paedophile (Level -3) . In this level, the subject has stopped talking to the bot. The strategy in this stage is to look for a affection in exchange for sex. We decided this strategy because a lot of paedophiles try to hide themselves to not get caught.*Possibly yes (Level +1). In this level, the subject shows interest inthe conversation and asks about personal topics. The topics of the bot arefavourite films, music, personal style, clothing, drugs and alcohol consumption and family issues. The bot is not too explicit in this stage.*Probably yes ( Level +2). In this level, the subject continues interested in the conversation and the topics become more private. Sex situations and experiences appear in the conversation and the bot does not avoid talking about them. The information is more detailed and private than before because we have to make the subject believe that he/she owns a lot of personal information for blackmailing. After reaching this level, it cannot decrease again.*Allegedly paedophile (Level +3). In this level, the system determines that the user is an actual paedophile. The conversations about sex becomes more explicit. Now, the objective is to keep the conversation active to gather as much information as possible. The information in this level is mostly sexual. The strategy in this stage is to give all the private information of the child simulated by the bot. After reaching this level, it cannot decrease again.
  12. When a new subject starts a conversation with Negobot the system is activated, and starts monitoring the input from the user. Besides, Negobot registers the conversations maintained with every user for future references, and to keepa record that could be sent to the authorities in case of determining that the subject is a paedophile.
  13. As youmayhave observestheconversations are in spanish, buttheretreivedconversationsfromperverted-justice, theoneswe use tofeedourknowledgesystem, are in english.Sincewedidn’thavespanishconversationsfrom real paedophiles, wedecidedtostoreourknowledge in English, and use on-line translationsystemstoadapttothatlanguage. In this case wetranslatedtheconversationfromspanishtoenglish, queriedthesystemtoknowiftheconversationisdisturbing, and then use thatknowledgetoreply back.
  14. First, despite current translation systems are good, they are far to be perfect. Therefore, the language is one of the most important issues. To solve it, we should obtain already classified conversations in other languages.Besides, the subsystem that adapts the way of speaking (i.e., child) should be improved. To this end, we will perform a further analysisof how young people speak on the Internet. Finally, there are some limitations regarding how the system determines the change of a topic. They are intrinsic to the language, and its solution is not simple
  15. And of course, wewill try toworkwiththeauthoritiestoadaptthissystemtotheirneeds. In thisproject, financedbytheBasqueGovernment, wehavehadthepossibilitytoworkwithaninternationalcontentfilteringorganisation and wethinkthatwecouldformaninterestingpartnershipwiththespanishcyber-crimeauthorities.