Papers with mining hard negatives
Generating Contrastive Narratives Using the Brownian Bridge Process for Narrative Coherence Learning (2024.acl-long)
Copied to clipboard
| Challenge: | Existing methods for learning narrative coherence are coarse-grained and superficial . existing methods are inadequate for learning negative samples, which are irrelevant or repetitive . |
| Approach: | They propose two strategies for mining hard negatives using the Brownian Bridge process . they evaluate the method on several tasks and show it is applicable to many applications . |
| Outcome: | The proposed method proves that it is applicable to many applications. |
On Synthetic Data Strategies for Domain-Specific Generative Retrieval (2025.acl-long)
Copied to clipboard
| Challenge: | Generative retrieval models can be used to generate ranked lists of potentially relevant document identifiers for a user query. |
| Approach: | They propose a synthetic data generation strategy for a two-stage training framework that focuses on learning to decode document identifiers from queries and a strategy for mining hard negatives based on initial model's predictions. |
| Outcome: | The proposed model can generate ranked lists of potentially relevant document identifiers for a user query and then refine ranking through preference learning. |