Filters








824,079 Hits in 5.4 sec

Exploration in two-stage recommender systems [article]

Jiri Hron and Karl Krauth and Michael I. Jordan and Niki Kilbertus
2020 arXiv   pre-print
Two-stage recommender systems are widely adopted in industry due to their scalability and maintainability.  ...  Modeled as a contextual bandit problem, we find LinUCB (a near optimal exploration strategy for single-stage systems) may lead to linear regret when deployed in two-stage recommenders.  ...  We plan to publish an extended version with additional experiments on real-world data and extended discussion in the near future.  ... 
arXiv:2009.08956v1 fatcat:eyoawdb2mferboe3xvy745vd7q

Two-Stage Neural Contextual Bandits for Personalised News Recommendation [article]

Mengyan Zhang, Thanh Nguyen-Tang, Fangzhao Wu, Zhenyu He, Xing Xie, Cheng Soon Ong
2022 arXiv   pre-print
Existing personalised news recommendation methods focus on exploiting user interests and ignores exploration in recommendation, which leads to biased feedback loops and hurt recommendation quality in the  ...  We propose a two-stage hierarchical topic-news deep contextual bandits framework to efficiently learn user preferences when there are many news items.  ...  Two-Stage Deep Recommendation Framework Recall our goal is to sequentially recommend m ≥ 1 news items to users in a large scale recommender system.  ... 
arXiv:2206.14648v1 fatcat:zztv4xkrvrcgvpez7uhn4icjea

Designing Explanation Interfaces for Transparency and Beyond

Chun-Hua Tsai, Peter Brusilovsky
2019 International Conference on Intelligent User Interfaces  
In this work-in-progress paper, we presented a participatory process of designing explanation interfaces for a social recommender system with multiple explanatory goals.  ...  We went through four stages to identify the key components of the recommendation model, expert mental model, user mental model, and target mental model.  ...  In this work-in-progress paper, we presented a stage-based participatory process [1] for integrating seven exploratory goals into real-world hybrid social recommender system.  ... 
dblp:conf/iui/TsaiB19a fatcat:357kg7ggcrc67d5dmqyuipwaya

Supporting Online Consumers by Identifying Consistency Distance Among Advice Sources

Hongki Kim, Izak Benbasat, Hasan Cavusoglu
2017 International Conference on Information Systems  
., not system-supported. In addition, the CDITs are costly and complicated for designers as well as consumers.  ...  better product selection decisions, and 2) which combination of a CDIT and information search stage is the most effective in improving decision performance.  ...  It is thus there are two applicable CDITs in the selecting a source stage and four combinations of applicable CDITs across the exploring a product list and elaborating a product stages.  ... 
dblp:conf/icis/KimBC17a fatcat:peq44zfljrbjdljp5g2hqnasq4

Recommender using Hybrid Approach for Candidate Generation

Gunjal Tejal H., Mutake Rupali N., Bhor Pranali S., Khatri Anand A.
2020 International journal of computer science and mobile computing  
A recommendation system helps user's find compelling content in a large corpus. Candidate generation is the first stage of recommendation. The system generates a set of appropriate candidates.  ...  Users are continuously come across with situations in which they have many options to choose and need help exploring or sorting down the possibilities.  ...  [3] In this paper neural networks and self-attention recommendation system is used. In this paper there are two stages of recommendation system. Stage-1.  ... 
doi:10.47760/ijcsmc.2020.v09i12.012 fatcat:dvazbqdfdjgchl4ef2jfdreqe4

A proactive recommendation system for writing

Mari Carmen Puerta Melguizo, Lou Boves, Anita Deshpande, Olga Muñoz Ramos
2007 Proceedings of the 14th European conference on Cognitive ergonomics invent! explore! - ECCE '07  
In two experiments we explore the effects of a PRS during these phases.  ...  A Proactive Recommender System (PRS) retrieves information relevant to the text being written, and presents it automatically.  ...  PROACTIVE RECOMMENDATION SYSTEMS FOR WRITING Proactive Recommendation Systems (PRSs) retrieve large quantities of documents, decide what available information is most likely relevant to the text being  ... 
doi:10.1145/1362550.1362569 dblp:conf/ecce/MelguizoBDR07 fatcat:kv7h4w637vb47iwad3om53se6u

A Theoretical Analysis of Two-Stage Recommendation for Cold-Start Collaborative Filtering [article]

Xiaoxue Zhao, Jun Wang
2016 arXiv   pre-print
In this paper, we study a simple two-stage recommendation combining a sequential and a batch solution together.  ...  the other hand, it is also necessary to allocate resources that are useful for learning the target's properties in order to recommend more relevant ones in the future.  ...  in the next stage -a property that can guide the system to find promising users in the next stage.  ... 
arXiv:1601.04745v1 fatcat:pswvl2noyjembokkare4kjzioq

A proactive recommendation system for writing: Helping without disrupting

Mari Carmen Puerta Melguizo, Lou Boves, Olga Muñoz Ramos
2009 International Journal of Industrial Ergonomics  
In two experiments we explore the effects of a PRS during these phases.  ...  A Proactive Recommender System (PRS) retrieves information relevant to the text being written, and presents it automatically.  ...  PROACTIVE RECOMMENDATION SYSTEMS FOR WRITING Proactive Recommendation Systems (PRSs) retrieve large quantities of documents, decide what available information is most likely relevant to the text being  ... 
doi:10.1016/j.ergon.2008.10.004 fatcat:rm2u3knidzhgbi7xoaib6mh3oy

Using Visualizations to Encourage Blind-Spot Exploration

Jayachithra Kumar, Nava Tintarev
2018 ACM Conference on Recommender Systems  
We compare the effectiveness of two visualizations -a bar-line chart and a scatterplot -for increasing a user's intention to explore new content.  ...  In this paper, we help users to better understand their consumption profiles by exposing them to their unexplored regions, thereby indirectly nudging them to diverse exploration.  ...  For ease of explanation, we divide our evaluation process into two conceptual stages: Stage 1 -where we evaluate user's understanding of visualizations, and Stage 2 -where we observe a user's music exploration  ... 
dblp:conf/recsys/KumarT18 fatcat:odcw5kynbzcjpgerkgaezuddsu

FashionNet: Personalized Outfit Recommendation with Deep Neural Network [article]

Tong He, Yang Hu
2018 arXiv   pre-print
To achieve personalized recommendation, we develop a two-stage training strategy, which uses the fine-tuning technique to transfer a general compatibility model to a model that embeds personal preference  ...  With the rapid growth of fashion-focused social networks and online shopping, intelligent fashion recommendation is now in great need.  ...  We therefore explore the use of deep networks for outfit recommendation.  ... 
arXiv:1810.02443v1 fatcat:hhbrgmeez5cl7df3mj6grhfqke

Pathway-Finder: An Interactive Recommender System for Supporting Personalized Care Pathways

Rui Liu, Raj Velamur Srinivasan, Kiyana Zolfaghar, Si-Chi Chin, Senjuti Basu Roy, Aftab Hasan, David Hazel
2014 2014 IEEE International Conference on Data Mining Workshop  
In this demonstration paper, we propose Pathway-Finder, an interactive recommender system to visually explore and discover clinical pathways.  ...  Additionally, the system implements a big-data infrastructure using Spark that is hosted as a HDinsight cluster on Microsoft Azure for Research platform to support real-time recommendation and visualization  ...  SYSTEM DEMONSTRATION In this section we demonstrate the four stages of Pathway-Finder: 1) Initial input collection; 2) Comorbidities exploration; 3) Intervention recommendation; 4) Outcome prediction and  ... 
doi:10.1109/icdmw.2014.37 dblp:conf/icdm/LiuSZCRHH14 fatcat:nves22rea5adjj5lovrtohjtje

Personalized pricing recommender system

Toshihiro Kamishima, Shotaro Akaho
2011 Proceedings of the 2nd International Workshop on Information Heterogeneity and Fusion in Recommender Systems - HetRec '11  
Based on the analysis of the experimental results, we reveal further issues in designing a personalized pricing recommender system.  ...  We then propose a method for adding price personalization to standard recommendation algorithms which utilize two types of customer data: preferential data and purchasing history.  ...  Multi-Stage Classification Our system identifies a type of a customer by the process shown in Figure 2 . This process consists of two major stages: the prescreening and main stages.  ... 
doi:10.1145/2039320.2039329 fatcat:kdefonswdvf5rc7lb5qasnrizy

RecPipe: Co-designing Models and Hardware to Jointly Optimize Recommendation Quality and Performance [article]

Udit Gupta, Samuel Hsia, Jeff Zhang, Mark Wilkening, Javin Pombra, Hsien-Hsin S. Lee, Gu-Yeon Wei, Carole-Jean Wu, David Brooks
2021 arXiv   pre-print
In particular, RPAccel processes queries in sub-batches to pipeline recommendation stages, implements dual static and dynamic embedding caches, a set of top-k filtering units, and a reconfigurable systolic  ...  Deep learning recommendation systems must provide high quality, personalized content under strict tail-latency targets and high system loads.  ...  We propose a new system, RecPipe, that enables design space exploration and optimization for multi-stage recommendation inference.  ... 
arXiv:2105.08820v2 fatcat:tsq6jygecvdo5l2bgs5pcbqbmu

The Effects of Cross-modal Collaboration on the Stages of Information Seeking

Dena Al-Thani, Tony Stockman, Anastasios Tombros
2015 Proceedings of the XVI International Conference on Human Computer Interaction - Interacción '15  
The findings showed that the different stages of the process were performed most of the time individually; however it was observed that some collaboration took place in the results exploration and management  ...  stages.  ...  In contrast, in systems that support explicit collaboration, users explicitly work together in the query formation and results exploration stages.  ... 
doi:10.1145/2829875.2829925 dblp:conf/interaccion/Al-ThaniST15 fatcat:6hdqge27dbhf5f54m55awndqqa

Adapt to Emotional Reactions in Context-aware Personalization

Yong Zheng
2016 ACM Conference on Recommender Systems  
However, there are no work exploring the effect of emotional reactions (or expressions) in the recommendation process.  ...  Users' emotions have been demonstrated as one of effective context information in recommender systems.  ...  Based on the introduction about the affective recommender systems [34] , the emotional information in three stages may be useful: entry stage (i.e., before the activity), consumption stage (i.e., during  ... 
dblp:conf/recsys/Zheng16 fatcat:rhd6nwmvqrae3jhzhtn34tn2ua
« Previous Showing results 1 — 15 out of 824,079 results