Recommendation Mining Minds Interpreter Service Curation Layer MMV-2.5
Overview 2 / • Personalization is a key element in Recommender Systems • Personalization consists of tailoring a service or a product to accommodate specific needs of individuals • Contextual Information combined with User Preferences enable Personalization • Recommendation Interpreter performs interpretation according to the contextual information and preferences of the user in order to deliver the appropriate recommendations at right time http://www.quora.com/What-is-the-definition-of-personalization
Goal and Objectives 3 / • Goal Physical Activities • Providing context-aware and personalized wellbeing recommendations Health Condition Preferences • Objectives Personalized Recommendations • Interpreting recommendations to address • Receptiveness of the user for recommendation • Preferences of the user for recommendation Time Profile • User friendly explanation of the Schedule recommendation
Motivation 4 / To deliver recommendation at appropriate time based on user current context To filter out unnecessary recommendations based on user preferences To explain recommendation according to situation for user engagement
Challenges and Solutions 5 / Motivations Challenges Solutions S1 To deliver recommendation User Receptivity Evaluation Context aware recommendations at appropriate time based (When to deliver?) on user current context S2 Filtering out unnecessary Exploiting user preferences recommendations based on User aware content filtration (What, whom and how to deliver?) user preferences S3 Multidimensional explanations Explaining recommendation Situation aware explanations according to situation (How to relate user’s surroundings?)
Conceptual View / 6 Deliver the recommendation according to present context and preferences Recommendation Interpreter Recommendation Is user receptive? Is Rec suitable? Rec User Status Yes Recommendations Filter and Weather is Rainy “take umbrella with you” Evaluation Builder Explain Rec User Educational Aid Location Weather Preferences IDB Emotion Location Explanation HLC Generation HLC special condition
7 Component Architecture / Recommendation Interpreter S1 S2 Context Interpreter Content Interpreter Content Filterer SNS Trend Identifier Context Context Selection Interpreter Select Trend Trend Filter Rec Alternative Selector Processor Context Moderator Recommendation Supporting Builder S3 Data Manager Explanation Manager Preparation Layer Results Situation Education Support Blocking Template Detection Rules UI/UX Resource Service Resource Explanation Linker Selector Orchestrator Generator Global Preferences DCL Lifelog Data
Complete Communication Workflow of Interpreter 8 / Uid Uid, Context, Preferences 7 4 Recommendation Interpreter 8 Context, Pref 10 Content Interpreter Context Interpreter Uid, Rec, Uid, Situation Recommendation Event Builder Orchestrator 3 9 2 Uid, Rec 11 Rec, Context 5 12 Explanation Manager Data Preparer Uid, Situation Event Context Uid Uid, 10a 8a 12a 14 Blocking Global Template Rules preferences 6 1 Uid, Personalized 13 13 Recommendation Uid, Personalized Recommendation DCL SL (UI/UX)
Execution Flow / Service Curation Layer Recommendation Interpreter Context Interpreter Content Interpreter SO receives situation event (SE) from 1 LLM Select Service Moderator Moderator Alternative Context Orchestrato 2 Select Selector Recv Rec. Context Eval. Context request is sent to SO r Path Receive 2 SCL Recv Context Process Vect. Context Final Rec Services 4 3 5 Pref-based EP. 1 User Status Eval. Context is received by Moderator Filter EP. 2 Fetch Blok Context Moderator sends context to Context 4 Filter Recommendation Rules Interpreter Match Rules Selector for evaluation Context Eval. Input/outpu t Adapter 5 Context Selector requests for the Filter Recom. Explanation Manager RB Data blocking rules Req/Resp. Generate Alternatives Moderator 3 Explanation Blocking rules are fetched from the 6 RI Data Pref-based Filter 1 Recv. Desc Generation Req/Resp. repository Recom. Eval. Desc Fetch Template Receive/Sen Process 6 d Interpretatio Template If loc : “Home” AND HLC= Education 6 n Support Post Proc. “Sleeping” Receive/Sen Blocked Global Template If HLC= “Having Meal” d Fetch URL Rules Pref If HLC = “Commuting” …
Execution Flow / Service Curation Layer Recommendation Interpreter Service Context Interpreter Content Interpreter Orchestrator Moderator Select Alternative Moderator Context Selector Receive Select Path SCL Services Recv Rec. Context Eval. Context EP. 1 Fetch Block Process Vect. Rules Recv Context EP. 2 7 Context Final Rec. Pref-based Interpreter User Status Filter Match Rules Input/output Eval. 9 Filter Recommendation 8 Adapter 8 return flag Match(rules, context) Gen. Explanation Manager RB Data Context Eval. { Alternatives Moderator Explanation Req/Resp. flag = -1 9 Generation rules.add(scanFile.nextLine()); Pref-based Recv. Desc User_Status_Eavl() { RI Data Filter Recom. if(rules.contains(context)){ Filter If flag==-1 Fetch Template Req/Resp. flag=1; Delay_Rec() Eval. Desc break; } else Recom. Both context and rules are forwarded 7 Repositories Process Content_Interpreter.Select_Path() return flag; Receive/Send to the Context Interpreter Template Education } } Interpretation Support 8 Context Interpreter evaluates context Blocked Global Receive/Send Fetch URL Post Proc. Templat against the rules and decides about User Rules Pref e availability Status If user is available then Content 9 Interpreter is invoked otherwise recommendation is delayed
Execution Flow / 14 Service Curation Layer Prepared Context Matrix 10 Select Rec Eval Recommendation Interpreter Contextual Matrix Path Content Interpreter Context/Rec Walking Running Stretching Cycling Sitting selt_path(Rec) { Out doors 1 1 1 1 1 Moderator Select Alternative Amusement 11 1 0 1 0 1 If (Rec.len == 1) Select Path Context Eval. Select_Alternative() Sunny 1 1 1 1 1 12 Else Happiness 1 1 1 1 1 Process Vect. Filter_Recommendation() 13 } Aggregate 1 0 1 1 1 Pref-based Final Rec 14 Walking Stretching Filter 10m 15m Select “ Select Alternative ” or “ Filter 10 Filter Recommendation Recommendation ” is called for further get_pref (user_id); Gen. processing Context Eval. Alternatives 11 Global preferences are fetched Pref-based Filter Recom. Filter Search Alternative Recommendation is 12 Repositories current Recommendation (running) is unsuitable 13 For multiple alternative recommendations, Blocked Global Template User’s Preferences are weighed in Rules Pref 14 “ Walking ” is preferred over “ Stretching ” by the user therefore “ Walking ” is treated as final recommendation
Execution Flow / Service Curation Layer Recommendation Interpreter Service Context Interpreter Content Interpreter Final Recommendation if forwarded 15 to Explanation Manager Orchestrator Moderator Select Alternative Context Selector Moderator Receive No explanatory sentence received SCL Services Recv Rec. Select Path Context Eval. 16 Context EP. 1 Fetch Block Explanation Generation component 17 Process Vect. Rules Recv Context is invoked EP. 2 Context A template is fetched from the local 18 Pref-based 15 Final Rec Interpreter repository and forwarded to further User Status Received Filter Match Rules Input/output 19 Eval. 15 processing Description Filter Recommendation Adapter Gen. Explanation Manager RB Data Context Eval. Alternatives get_Description() Moderator Explanation Req/Resp. //empty string Generation Pref-based Recv. Desc 16 RI Data Filter Recom. Filter Forward Fetch Template Req/Resp. 16 18 Description 17 17 19 Eval. Desc Recom. Repositories Process Evaluate Description Receive/Send Template Education forward_Descption(); Interpretation Support Eval_Desc() { Blocked Global Fetch URL Post Proc. Receive/Send Template if (Descprption.isEmpty()) Rules Pref Explanation_Generation() else Education_Support() }
Recommend
More recommend
