A5027396594- Topic Modeling
- Recommender Systems and Techniques
- Speech and dialogue systems
- Web Data Mining and Analysis
- Information Retrieval and Search Behavior
- Human-Animal Interaction Studies
- Expert finding and Q&A systems
- Auction Theory and Applications
- Advanced Bandit Algorithms Research
- Biodiesel Production and Applications
- Mobile Crowdsensing and Crowdsourcing
- Natural Language Processing Techniques
- Animal Behavior and Welfare Studies
- Data Quality and Management
- Industrial Technology and Control Systems
- Privacy-Preserving Technologies in Data
- Explainable Artificial Intelligence (XAI)
- Sentiment Analysis and Opinion Mining
- Text Readability and Simplification
- Evolutionary Psychology and Human Behavior
The University of Queensland
2024
Shanghai Jiao Tong University
2023-2024
With the emergence of Large Language Models (LLMs), there has been a significant improvement in programming capabilities models, attracting growing attention from researchers. Evaluating LLMs is crucial as it reflects multifaceted abilities LLMs, and numerous downstream applications. In this paper, we propose CodeApex, bilingual benchmark dataset focusing on comprehension, code generation, correction LLMs. Programming comprehension task tests multiple-choice exam questions covering...
Multimodal Recommendation focuses mainly on how to effectively integrate behavior and multimodal information in the recommendation task. Previous works suffer from two major issues. Firstly, training process tightly couples module by jointly optimizing them using sharing model parameters, which leads suboptimal performance since signals modality often provide opposite guidance for parameters updates. Secondly, previous approaches fail take into account significant distribution differences...
Integrated ranking is a critical component in industrial recommendation platforms. It combines candidate lists from different upstream channels or sources and ranks them into an integrated list, which will be exposed to users. During this process, take responsibility for channel providers, the system needs consider exposure fairness among channels, directly affects opportunities of being displayed Besides, personalization also requires user's diverse preference on besides items. Existing...
With the development of recommender systems, it becomes an increasingly common need to mix multiple item sequences from different sources. Therefore, integrated ranking stage is proposed be responsible for this task with re-ranking models. However, existing methods ignore relation between sequences, thus resulting in local optimum over interaction session. To resolve challenge, paper, we propose a new model named NFIRank (News Feed Integrated Ranking reinforcement learning) and formulate...
The essence of information retrieval (IR) is to find the most useful items (or documents) according user's need and present users in form a ranking list. widely used evaluation metrics for list are NDCG, MAP, hit ratio etc., which based on strong assumptions users' examining click behaviors when interacting with In modern IR scenarios, it has been shown that behavior can be highly personalized, diverse dynamic, leads failure those assumptions. Click models (CMs) proposed learn such complex...