Information market based recommender systems fusion Efthimios - PowerPoint PPT Presentation

Information Management Unit / ICCS of NTUA imu.iccs.gr Information market based recommender systems fusion Efthimios Bothos Konstantinos Christidis Dimitris Apostolou National Technical National Technical University of Piraeus University of Athens University of Athens Greece Greece Greece Gregoris Mentzas National Technical University of Athens Greece Information Management Unit, HetRec Workshop, RecSys October 23-27, 2011 Chicago, IL USA 1

Information Management Unit / ICCS of NTUA imu.iccs.gr Outline  Introduction  Background on Information Markets  Information market based recommender systems fusion  Experiments and Results  Conclusions and Further Work Information Management Unit, HetRec Workshop, RecSys October 23-27, 2011 Chicago, IL USA 2

Information Management Unit / ICCS of NTUA imu.iccs.gr Research Area: Ensemble Recommenders  Recent work in recommender systems proposes the acquisition of results by combining sets of recommendation models  Aggregation of the predictions of different base algorithms - the ensemble - to obtain a final prediction  The combination of different predictions into a final prediction is also referred to as blending or fusion  E.g. the models proposed for the NetFlix Prize have been combined to address the problem of recommendation in the specific dataset  Most existing methods presuppose restrictive assumptions  Most ensembles of recommender s have a constant composition  The training data performance has to be a good proxy for subsequent actual performance  Cannot easily adapt to changes in future user behavior Information Management Unit, HetRec Workshop, RecSys October 23-27, 2011 Chicago, IL USA 3

Information Management Unit / ICCS of NTUA imu.iccs.gr Our Approach  In this paper we use the market paradigm for blending recommender systems  Market participants are computational agents  Representing different base recommenders  Agents invest/bet on the recommendation they foresee to be correct  Based on the information provided by their corresponding base recommender  The recommendation is based on the market outcome  It depends on the wealth of the participants and reflects the „wealth - weighted opinions' of the base recommenders Information Management Unit, HetRec Workshop, RecSys October 23-27, 2011 Chicago, IL USA 4

Information Management Unit / ICCS of NTUA imu.iccs.gr Outline  Introduction  Background on Information markets  Information market based recommender systems fusion  Experiments and Results  Conclusions and Further Work Information Management Unit, HetRec Workshop, RecSys October 23-27, 2011 Chicago, IL USA 5

Information Management Unit / ICCS of NTUA imu.iccs.gr Information Markets (IMs) In general markets provide mechanisms for risk sharing or resource  allocation Market prices are able to aggregate and convey information  Efficient markets hypothesis: market in which prices always “fully reflect”  available information is called “efficient” (Fama, 1970) IMs are markets designed and run for the primary purpose of mining and  aggregating information scattered among participants Also known as Prediction markets, Decision markets  IMs make use of specifically designed contracts that yield payments  based on the outcome of uncertain future events  Contrary to traditional equity markets contracts are not tied to a claim of an ownership stake in a firm  The assets are claims that will pay off an amount which depends upon the state of the world Information Management Unit, HetRec Workshop, RecSys October 23-27, 2011 Chicago, IL USA 6

Information Management Unit / ICCS of NTUA imu.iccs.gr An Example Will be announced today, but since price reflects probability we are pretty confident the GDP will be positive! Will United States GDP growth for Q3 of 2011 be positive?  Traders who believe it will be positive buy contracts, otherwise sell contracts Information Management Unit, HetRec Workshop, RecSys October 23-27, 2011 Chicago, IL USA 7

Information Management Unit / ICCS of NTUA imu.iccs.gr IMs Applications IMs with human participants have done well in various contexts:   Predictions  Iowa Electronic Markets beat US presidential election polls 451/596 (Berg et al. 2008)  Oscar winners: Hollywood Stock Exchange beats individual and average forecasts of 5 experts (Lamare, 2007)  NFL: Markets rank 11 th and 12 th against 1947 humans (human average 39th) (Servan-Schreiber et al. 2005)  Decision support (opinion polling)  Hewlett-Packard market beats official forecasts in 6 out of 8 events (Chen & Plott 2002)  Preferences  Selection of product concepts with IAMs provides similar results with surveys (Dahan et al. 2010) IMs with computational agents have also done well:   eg. Predicting the Oscar awards (Bothos et al. 2010) Information Management Unit, HetRec Workshop, RecSys October 23-27, 2011 Chicago, IL USA 8

Information Management Unit / ICCS of NTUA imu.iccs.gr Design Elements  Contracts reveal participants expectations regarding the status of future events  Payoff depends on the outcome of future events  E.g. a futures contract pays $1 if the event occurs, nothing otherwise  Exchange medium  Can be either real or play money  Trading Mechanisms  Define the rules of the market, which specify how orders are placed and how the price changes Information Management Unit, HetRec Workshop, RecSys October 23-27, 2011 Chicago, IL USA 9

Information Management Unit / ICCS of NTUA imu.iccs.gr Design Elements  Participants bring liquidity and generate the efficient price  This means that even uninformed traders must participate. When only rational traders participate, the “No Trading Theorem” (Milgrom and Stokey, 1982) effect appears and the market cannot function.  Participants should be (Surowiecki, 2004)  Diverse so that people offer different pieces of information  Independent , so that participants pay attention mostly to their own information, and do not worry about what others think  De-centralized , so that no one at the top is dictating the crowd's answer  Openness with respect to information Information Management Unit, HetRec Workshop, RecSys October 23-27, 2011 Chicago, IL USA 10

Information Management Unit / ICCS of NTUA imu.iccs.gr Outline  Introduction  Background on Information markets  Information market based recommender systems fusion  Experiments and Results  Conclusions and Further Work Information Management Unit, HetRec Workshop, RecSys October 23-27, 2011 Chicago, IL USA 11

Information Management Unit / ICCS of NTUA imu.iccs.gr Our methodology We consider an IM as an ensemble recommender Trained Recommenders  Acting as Agents which can potentially be employed in any recommendation  problem For a given dataset a set of base  Agent1 Rec1 recommenders is considered Unrated Item Base recommenders are trained  An IM is composed where its participants are  Information Agent2 Rec2 computational agents representing different Predicted Market rating base recommenders Agents invest on the rating option they  Agent3 Rec3 foresee to be correct They make use of information provided by  their corresponding base recommender When the actual rating is revealed, the agents  that predicted correctly are rewarded AgentN RecN Recommendation Information Management Unit, HetRec Workshop, RecSys October 23-27, 2011 Chicago, IL USA 12

Information Management Unit / ICCS of NTUA imu.iccs.gr Agents Our agents follow the Belief-Desire-Intention design paradigm   According to the BDI framework an agent is characterized by its beliefs, goals (desires), and intentions  The agent intends to achieve his goals given his beliefs about the world Belief : Stems from the base recommenders  Desire : To maximize their wealth   Wealth is determined by the agent‟s forecasting accuracy Intention : Betting function which defines the wealth percentage an agent  will allocate for each rating option  Depends on the item to be rated  We use a constant betting function: Agents invest independent of the market price N: number of possible options m the agent / recommender m estimate m the otput of the base recommender m Weatlth m :the wealth of agent m Information Management Unit, HetRec Workshop, RecSys October 23-27, 2011 Chicago, IL USA 13

Information Management Unit / ICCS of NTUA imu.iccs.gr Information Market Zero-sum game i.e. the total amount of money collectively owned by the  participants is conserved after each new item is presented Prices denote probabilities of an outcome being correct and sum up to one  Equilibrium price : A unique price that satisfies  Total wealth conservation and  For constant betting functions solution is provided as follows (Barbu‟10, ICML)  k denotes the rating option, x the output of the base recommender, m the agent Rewards : The agent is rewarded based on the investment he made on the correct  outcome Information Management Unit, HetRec Workshop, RecSys October 23-27, 2011 Chicago, IL USA 14