A5002329426- Recommender Systems and Techniques
- Topic Modeling
- Advanced Graph Neural Networks
- Explainable Artificial Intelligence (XAI)
- Image Retrieval and Classification Techniques
- Data Mining Algorithms and Applications
- Advanced Bandit Algorithms Research
- Sentiment Analysis and Opinion Mining
- Video Analysis and Summarization
Universidade de São Paulo
2020-2023
Brazilian Society of Computational and Applied Mathematics
2020-2022
Regarding recommender systems, there has been a historically dominance on the literature in favor of collaborative algorithms over content-based ones. However, latter can work better with users applications and be more transparent. Therefore, this paper, we propose WordRecommender, an explainable algorithm that calculates similarity by semantic proximity. On its preprocessing step, reviews movie context are analyzed to obtain aspects, defined as words high sentimental value. After that,...
Recommendations engines use interactions between users and items to predict the preferences of generate recommendations for them. However, because they rely on historical data, user’s interest at moment may not be captured. In this context, Conversational Recommender Systems (CRSs) have been proposed in order provide that suggestions based current interests by eliciting information from user turns which system can ask understand more about time or recommend. regard, we propose CRSs...
Explanations in recommender systems are essential improving trust, transparency, and persuasion. Recently, using Knowledge Graphs (KG) to generate explanations gained attention due the semantic representation of information which items their attributes represented as nodes, connected by edges, representing connections among them. Model-agnostic KG explainable algorithms can be based on syntactic approaches or graph embeddings. The impact embedding strategies generating meaningful still needs...
Abstract Most studies on recommender systems focus collaborative algorithm approaches over content‐based recommendation due to their better accuracy results. However, the advantage of latter is that it more effective and transparent with user applications. This article proposes WordRecommender, an explainable calculates similarity by semantic proximity. Its preprocessing step involves analyses movie reviews obtain aspects, defined as relevant words high sentimental value. Recommendations are...
Explanations are crucial for improving users' transparency, persuasiveness, engagement, and trust in Recommender Systems (RSs). However, evaluating the effectiveness of explanation algorithms regarding those goals remains challenging due to existing offline metrics' limitations. This paper introduces new metrics evaluation validation based on items properties used form sentence an explanation. Towards validating metrics, results three state-of-the-art post-hoc were evaluated six RSs,...