Papers by Yacine Kessaci
Towards a more Robust Evaluation for Conversational Question Answering (2021.acl-short)
Copied to clipboard
| Challenge: | Conversational Question Answering (CQA) is a new form of NLP . it uses conversation history to extract the answer of the current question. |
| Approach: | They propose to use conversation history to evaluate models which can access the ground truth answers of previous turns at each turn of the conversation. |
| Outcome: | The proposed evaluation protocol severely limits the effectiveness of the proposed models in fully autonomous chatbots and leads to unsuspected biases in their behavior. |
Towards Multi-Document Question Answering in Scientific Literature: Pipeline, Dataset, and Evaluation (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Existing QA systems do not strictly enforce cross-document synthesis or exploit the explicit inter-paper structure that links sources. |
| Approach: | They propose a pipeline methodology for constructing a multi-document academic QA dataset . they detect communities based on citation networks and leverage Large Language Models . |
| Outcome: | The proposed method generates QA pairs related to multi-document content automatically and forms coherent communities based on citation networks and large language models. |
Structure-Aware Quantized Retrieval for Long-Document Question Answering (2026.findings-acl)
Copied to clipboard
| Challenge: | Traditional long-document QA/RAG pipelines suffer from context fragmentation . Graph Neural Networks (GNNs) can encode hierarchy, but deep GNNs tend to over-smooth representations. |
| Approach: | They propose a framework that aligns hierarchical graph representations with a universal token vocabulary and integrates explicit structure into retrieval. |
| Outcome: | The proposed framework captures universal hierarchical patterns rather than overfitting to specific layouts. |