Papers by Fenia Christopoulou
A Walk-based Model on Entity Graphs for Relation Extraction (P18-2)
Copied to clipboard
| Challenge: | Existing models treat each relation in a sentence individually, but a graph-based model needs to consider multiple relations between entities to model the dependencies among them. |
| Approach: | They propose a graph-based neural network model that treats multiple pairs in a sentence simultaneously and considers interactions among them. |
| Outcome: | The proposed model performs comparable to the state-of-the-art systems on the ACE 2005 dataset without external tools. |
Inter-sentence Relation Extraction with Document-level Graph Convolutional Neural Network (P19-1)
Copied to clipboard
| Challenge: | Existing methods for inter-sentence relation extraction do not fully exploit such dependencies. |
| Approach: | They propose a model that captures local and non-local dependencies using multi-instance learning and bi-affine pairwise scoring to predict the relation of an entity pair. |
| Outcome: | The proposed model performs comparable to state-of-the-art models on biochemistry datasets. |
Training Dynamics for Curriculum Learning: A Study on Monolingual and Cross-lingual NLU (2022.emnlp-main)
Copied to clipboard
| Challenge: | Current approaches for NLU use CL to improve in-distribution data performance via heuristic-oriented or task-agnostic difficulties. |
| Approach: | They propose to use CL to improve in-distribution data performance by taking advantage of training dynamics as difficulty metrics instead of heuristic-oriented or task-agnostic difficulties. |
| Outcome: | The proposed model schedulers improve on in-distribution, out-of-distortion and zero-shot cross-lingual transfer datasets while being 20% faster on average. |
Connecting the Dots: Document-level Neural Relation Extraction with Edge-oriented Graphs (D19-1)
Copied to clipboard
| Challenge: | Existing approaches to document-level relation extraction use nodes and edges as relations between nodes. |
| Approach: | They propose an edge-oriented graph neural model for document-level relation extraction that uses different types of nodes and edges to create a document-based graph. |
| Outcome: | The proposed model can learn intra- and inter-sentence relations using multi-instance learning internally. |
Distantly Supervised Relation Extraction with Sentence Reconstruction and Knowledge Base Priors (2021.naacl-main)
Copied to clipboard
| Challenge: | Existing methods to facilitate distantly supervised relation extraction are noisy instances, long-tail relations and unbalanced bag sizes. |
| Approach: | They propose a multi-task approach to facilitate distantly supervised relation extraction by bringing closer the representations of sentences that contain the same Knowledge Base pairs. |
| Outcome: | The proposed approach improves performance on two datasets created via distant supervision. |
EntityCS: Improving Zero-Shot Cross-lingual Transfer with Entity-Centric Code Switching (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods for CS use dictionaries or parallel sentences with word-alignment to generate CS data by randomly switching words in a sentence. |
| Approach: | They propose a method that focuses on Entity-level Code-Switching to capture fine-grained cross-lingual semantics without corrupting syntax. |
| Outcome: | The proposed method captures fine-grained cross-lingual semantics without corrupting syntax. |
Text-to-Code Generation with Modality-relative Pre-training (2024.eacl-long)
Copied to clipboard
| Challenge: | Large pre-trained language models have been applied to programming language tasks with great success, often through further pre-training of a strictly-natural language model. |
| Approach: | They propose to map programming language modalities into the same embedding space by separating embeddable spaces between modality and modality-relative training objectives. |
| Outcome: | The proposed model can be adapted and represented differently depending on which modality they belong to and to the ultimate benefit of the downstream task. |
SparsePO: Controlling Preference Alignment of LLMs via Sparse Token Masks (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Current direct preference optimization algorithms focus on a strict set of tokens contributing signals of KL divergence and rewards to the loss function. |
| Approach: | They propose a method that automatically learns to weight the KL divergence and reward corresponding to each token during PO training. |
| Outcome: | The proposed method achieves +10% and +3% win-rate points in two PO scenarios. |