An Efficient Rehearsal Scheme for Catastrophic Forgetting Mitigation during Multi-stage Fine-tuning (2025.findings-naacl)
Copied to clipboard
| Challenge: | Existing approaches to fine-tune foundational models on new tasks or domains are costly and time-consuming. |
| Approach: | They propose a sampling scheme that prioritizes rehearsal of "collateral damage" samples . the scheme is computationally efficient and easy to implement, they say . |
| Outcome: | a new approach prioritizes rehearsal of “collateral damage” samples outperforms other continual learning methods. |
Similar Papers
The Best of Both Worlds: Combining Parallel and Sequential Inference Scaling via Aggregation Fine-Tuning (2026.findings-acl)
Copied to clipboard
| Challenge: | Empirical results show that AFT-trained models achieve substantial gains with test-time scaling. |
| Approach: | They introduce a supervised fine-tuning paradigm where models synthesize multiple draft responses into a single, refined answer. |
| Outcome: | Empirical results show that AFT-trained models outperform baseline models while eliminating external guidance. |
Recall and Learn: Fine-tuning Deep Pretrained Language Models with Less Forgetting (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods to fine-tune deep pretrained language models face catastrophic forgetting problems. |
| Approach: | They propose a recall and learn mechanism which integrates pretraining and downstream tasks into a single mechanism. |
| Outcome: | The proposed method achieves state-of-the-art performance on the GLUE benchmark and better average performance than directly fine-tuning of BERT-large. |
Mitigating Catastrophic Forgetting in Large Language Models with Self-Synthesized Rehearsal (2024.acl-long)
Copied to clipboard
Jianheng Huang, Leyang Cui, Ante Wang, Chengyi Yang, Xinting Liao, Linfeng Song, Junfeng Yao, Jinsong Su
| Challenge: | Existing methods to train LLMs on previous training data are not feasible in real-world applications because of catastrophic forgetting. |
| Approach: | They propose a framework that uses the LLM to generate synthetic instances for rehearsal and refine the instance outputs based on the synthetic inputs. |
| Outcome: | The proposed framework achieves superior or comparable performance compared to conventional rehearsal-based approaches while being more data-efficient. |
Improving Scheduled Sampling with Elastic Weight Consolidation for Neural Machine Translation (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Autoregressive models trained with maximum likelihood estimation suffer from exposure bias, i.e. the discrepancy between ground-truth prefixes used during training and model-generated prefix at inference time. |
| Approach: | They propose to use Elastic Weight Consolidation to better balance mitigating exposure bias with retaining performance. |
| Outcome: | The proposed method significantly outperforms maximum likelihood estimation and scheduled sampling baselines on four translation datasets. |
Temporal Sampling for Forgotten Reasoning in LLMs (2026.acl-long)
Copied to clipboard
Yuetai Li, Zhangchen Xu, Fengqing Jiang, Bhaskar Ramasubramanian, Luyao Niu, Bill Yuchen Lin, Xiang Yue, Radha Poovendran
| Challenge: | a new metric measures the percentage of questions that were answered incorrectly during fine-tuning . |
| Approach: | They propose a decoding strategy that draws outputs from multiple checkpoints along the training trajectory. |
| Outcome: | The proposed method improves reasoning performance and consistency across benchmarks. |
Revisiting Catastrophic Forgetting in Large Language Model Tuning (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Catastrophic Forgetting (CF) compromises the effectiveness of large language models during fine-tuning, yet the underlying causes of CF remain largely unexplored. |
| Approach: | They propose a method to flatten the model loss landscape to mitigate CF by flattening the loss landscape. |
| Outcome: | The proposed method complements existing anti-forgetting strategies, further enhancing the resistance of LLMs to CF. |
How to Fine-Tune Safely on a Budget: Model Adaptation Using Minimal Resources (2025.emnlp-industry)
Copied to clipboard
Anh C. Pham, Mihir Thalanki, Michael Sun, Aditya Chaloo, Ankita Gupta, Tian Xia, Aditya Mate, Ehi Nosakhare, Soundararajan Srinivasan
| Challenge: | Existing methods for fine-tuning safety examples are underdeveloped. |
| Approach: | They hypothesize that the effectiveness of a safety example is governed by its instruction-response behavior and its semantic diversity across harm categories. |
| Outcome: | The proposed method reduces harmfulness by up to 41% while adding only 0.05% more data to the fine-tuning set. |
Optimizing Lifelong Fine-Tuning for Multiple Tasks via Dataless Distribution Replay (2025.coling-main)
Copied to clipboard
| Challenge: | Existing methods to fine-tune large language models with minimal instruction data are prone to catastrophic forgetting during life-long fine- tuning. |
| Approach: | They propose a dataless distribution replay approach for life-long fine-tuning that replays the output of the linear layers at previous task stages. |
| Outcome: | The proposed method achieves significant improve- ments compared to several strong lifelong fine- tuning methods. |
Fine-tuning Large Language Models with Limited Data: A Survey and Practical Guide (2026.tacl-1)
Copied to clipboard
| Challenge: | Pre-trained language models provide strong foundations, but effective adaptation under data scarcity requires efficient and efficient fine-tuning techniques. |
| Approach: | They propose to review parameter-efficient fine-tuning techniques that lower training and deployment costs and domain and cross-lingual adaptation methods for both encoder and decoder models. |
| Outcome: | The proposed techniques lower training and deployment costs, domain and cross-lingual adaptation methods, and model specialization strategies. |
An Empirical Investigation Towards Efficient Multi-Domain Language Model Pre-training (2020.emnlp-main)
Copied to clipboard
| Challenge: | Pre-training large language models is a standard practice in the natural language processing community. |
| Approach: | They propose to use elastic weight consolidation to mitigate catastrophic forgetting when pre-trained large language models are evaluated on generic benchmarks. |
| Outcome: | The proposed model achieves state-of-the-art on out-of domain tasks with minimal pre-training . elastic weight consolidation provides best overall scores yielding only a 0.33% drop in performance across seven generic tasks while remaining competitive in bio-medical tasks. |