A5081007833- Topic Modeling
- Natural Language Processing Techniques
- Consumer Market Behavior and Pricing
- Auction Theory and Applications
- Speech Recognition and Synthesis
- Transportation and Mobility Innovations
- Expert finding and Q&A systems
- Advanced Bandit Algorithms Research
- Business Process Modeling and Analysis
- Semantic Web and Ontologies
- Multimodal Machine Learning Applications
- Service-Oriented Architecture and Web Services
- Speech and dialogue systems
Retrieval Augmented Generation (RAG) has become prevalent in question-answering (QA) tasks due to its ability of utilizing search engine enhance the quality long-form (LFQA). Despite emergence various open source methods and web-enhanced commercial systems such as Bing Chat, two critical problems remain unsolved, i.e., lack factuality clear logic generated answers. In this paper, we remedy these issues via a systematic study on answer generation LFQA. Specifically, first propose novel...
The prevailing issue of factual inconsistency errors in conventional Retrieval Augmented Generation (RAG) motivates the study Factual Consistency Evaluation (FCE). Despite various FCE methods proposed earlier, these are evaluated on datasets generated by specific Large Language Models (LLMs). Without a comprehensive benchmark, it remains unexplored how perform other LLMs with different error distributions or even unseen types, as may fail to detect types LLMs. To fill this gap, paper, we...
We study the budget allocation problem in online marketing campaigns that utilize previously collected offline data. first discuss long-term effect of optimizing decisions setting. To overcome challenge, we propose a novel game-theoretic value-based reinforcement learning method using mixed policies. The proposed reduces need to store infinitely many policies previous methods only constantly policies, which achieves nearly optimal policy efficiency, making it practical and favorable for...
The prevailing issue of factual inconsistency errors in conventional Retrieval Augmented Generation (RAG) motivates the study Factual Consistency Evaluation (FCE). Despite various FCE methods proposed earlier, these are evaluated on datasets generated by specific Large Language Models (LLMs). Without a comprehensive benchmark, it remains unexplored how perform other LLMs with different error distributions or even unseen types, as may fail to detect types LLMs. To fill this gap, paper, we...
We study the budget allocation problem in online marketing campaigns that utilize previously collected offline data. first discuss long-term effect of optimizing decisions setting. To overcome challenge, we propose a novel game-theoretic value-based reinforcement learning method using mixed policies. The proposed reduces need to store infinitely many policies previous methods only constantly policies, which achieves nearly optimal policy efficiency, making it practical and favorable for...
Aligning language models to human expectations, e.g., being helpful and harmless, has become a pressing challenge for large models. A typical alignment procedure consists of supervised fine-tuning preference learning. Most learning methods, such as RLHF DPO, depend on pairwise data, which inadequately address scenarios where feedback is point-wise, leading potential information loss suboptimal performance. Addressing this gap, we introduce Point-wise Direct Preference Optimization, novel...