A5101875685- Advanced Image and Video Retrieval Techniques
- Domain Adaptation and Few-Shot Learning
- Topic Modeling
- Multimodal Machine Learning Applications
- Advanced Graph Neural Networks
- Recommender Systems and Techniques
- Adaptive Dynamic Programming Control
- Distributed Control Multi-Agent Systems
- Adaptive Control of Nonlinear Systems
- Metaheuristic Optimization Algorithms Research
Northeastern University
2023-2024
Synthetic Biologics (United States)
2020
This paper proposes Text mAtching based SequenTial rEcommenda-tion model (TASTE), which maps items and users in an embedding space recommends by matching their text representations. TASTE verbalizes user-item interactions using identifiers attributes of items. To better characterize user behaviors, additionally attention sparsity method, enables to longer reducing the self-attention computations during encoding. Our experiments show that outperforms state-of-the-art methods on widely used...
Retrieval-Augmented Generation (RAG) has proven its effectiveness in mitigating hallucinations Large Language Models (LLMs) by retrieving knowledge from external resources. To adapt LLMs for RAG pipelines, current approaches use instruction tuning to optimize LLMs, improving their ability utilize retrieved knowledge. This supervised fine-tuning (SFT) approach focuses on equipping handle diverse tasks using different instructions. However, it trains modules overfit training signals and...
This paper proposes Text mAtching based SequenTial rEcommendation model (TASTE), which maps items and users in an embedding space recommends by matching their text representations. TASTE verbalizes user-item interactions using identifiers attributes of items. To better characterize user behaviors, additionally attention sparsity method, enables to longer reducing the self-attention computations during encoding. Our experiments show that outperforms state-of-the-art methods on widely used...
This paper proposes Multi-modAl Retrieval model via Visual modulE pLugin (MARVEL), which learns an embedding space for queries and multi-modal documents to conduct retrieval. MARVEL encodes with a unified encoder model, helps alleviate the modality gap between images texts. Specifically, we enable image understanding ability of well-trained dense retriever, T5-ANCE, by incorporating visual module's encoded features as its inputs. To facilitate retrieval tasks, build ClueWeb22-MM dataset...