Papers by Giannis Karamanolakis
MergeME: Model Merging Techniques for Homogeneous and Heterogeneous MoEs (2025.naacl-long)
Copied to clipboard
Yuhang Zhou, Giannis Karamanolakis, Victor Soto, Anna Rumshisky, Mayank Kulkarni, Furong Huang, Wei Ai, Jianhua Lu
| Challenge: | State-of-the-art methods for merging expert models with different architectures do not address parameter interference and require extensive fine-tuning to restore performance. |
| Approach: | They propose a method for merging experts with different architectures into a unified Mixture-of-Experts model with a goal of enhancing performance in each domain while retaining effectiveness on general tasks. |
| Outcome: | Experiments across multiple domains show that the proposed methods reduce fine-tuning costs and improve performance over state-of-the-art methods. |
WALNUT: A Benchmark on Semi-weakly Supervised Learning for Natural Language Understanding (2022.naacl-main)
Copied to clipboard
| Challenge: | Existing studies on weak supervision for NLU focus on a specific task or simulate weak supervision signals from ground-truth labels. |
| Approach: | They propose a benchmark to advocate and facilitate research on weak supervision for NLU . they use document-level and token-level prediction tasks as examples . |
| Outcome: | The proposed benchmark advocates and facilitates research on weak supervision for NLU tasks. |
Self-Training with Weak Supervision (2021.naacl-main)
Copied to clipboard
| Challenge: | State-of-the-art deep neural networks require large amounts of labeled training data that is expensive to obtain or not available for many tasks. |
| Approach: | They propose a weak supervision framework that leverages all available data for a given task . they leverage task-specific unlabeled data through self-training with a model that predicts pseudo-labels for instances that may not be covered by weak rules . |
| Outcome: | The proposed framework improves on state-of-the-art datasets on six benchmark tasks. |
TXtract: Taxonomy-Aware Knowledge Extraction for Thousands of Product Categories (2020.acl-main)
Copied to clipboard
| Challenge: | State-of-the-art methods for knowledge extraction are designed for a single category of product, but do not apply to real-life e-Commerce scenarios. |
| Approach: | They propose a taxonomy-aware knowledge extraction model that applies to thousands of categories organized in a hierarchical taxonomies. |
| Outcome: | The proposed model outperforms state-of-the-art methods on 4,000 categories in F1 and 15% across all categories. |
Cross-Lingual Text Classification with Minimal Resources by Transferring a Sparse Teacher (2020.findings-emnlp)
Copied to clipboard
| Challenge: | Existing approaches for transferring supervision across languages require expensive cross-lingual resources. |
| Approach: | They propose a cross-lingual teacher-student method that generates "weak" supervision in a target language using minimal cross-linguistic resources. |
| Outcome: | The proposed method outperforms state-of-the-art methods with a student classifier in 18 languages . it extracts and transfers only the most important task-specific seed words across languages based on translated seed words . |
Super-NaturalInstructions: Generalization via Declarative Instructions on 1600+ NLP Tasks (2022.emnlp-main)
Copied to clipboard
Yizhong Wang, Swaroop Mishra, Pegah Alipoormolabashi, Yeganeh Kordi, Amirreza Mirzaei, Atharva Naik, Arjun Ashok, Arut Selvan Dhanasekaran, Anjana Arunkumar, David Stap, Eshaan Pathak, Giannis Karamanolakis, Haizhi Lai, Ishan Purohit, Ishani Mondal, Jacob Anderson, Kirby Kuznia, Krima Doshi, Kuntal Kumar Pal, Maitreya Patel, Mehrad Moradshahi, Mihir Parmar, Mirali Purohit, Neeraj Varshney, Phani Rohitha Kaza, Pulkit Verma, Ravsehaj Singh Puri, Rushang Karia, Savan Doshi, Shailaja Keyur Sampat, Siddhartha Mishra, Sujan Reddy A, Sumanta Patro, Tanay Dixit, Xudong Shen
| Challenge: | a benchmark of 1,616 diverse NLP tasks and their expert-written instructions is used to test generalization of models to unseen tasks . a recent study shows that instruction-following models outperform instruction-based models by over 9% . |
| Approach: | They build a benchmark of 1,616 diverse NLP tasks and their expert-written instructions. |
| Outcome: | The proposed model outperforms existing instruction-following models by over 9% on the benchmark despite being smaller. |
Leveraging Just a Few Keywords for Fine-Grained Aspect Detection Through Weakly Supervised Co-Training (D19-1)
Copied to clipboard
| Challenge: | Current weakly supervised approaches for learning aspect classifiers require many fine-grained aspect labels, which are labor-intensive to obtain. |
| Approach: | They propose a weakly supervised approach that leverages seed words for aspect detection . they propose supervised student-teacher approach that uses teacher to train student models . |
| Outcome: | The proposed approach outperforms previous weakly supervised approaches by 14.1 F1 points on average in six domains of product reviews and six multilingual datasets of restaurant reviews. |
Weakly Supervised Attention Networks for Fine-Grained Opinion Mining and Public Health (D19-55)
Copied to clipboard
| Challenge: | Existing weakly supervised learning frameworks are used for segment classification . lack of segment labels prevents the use of standard supervised methods . |
| Approach: | They propose a model that uses weak supervision to train supervised models for segment-level classification . they propose sigmoid attention mechanism-based aggregation function to improve the model . |
| Outcome: | The proposed model outperforms state-of-the-art models for segment-level sentiment classification by 9.8% in F1 . |