Papers with CIT
Towards Better Generalization in Open-Domain Question Answering by Mitigating Context Memorization (2024.findings-naacl)
Copied to clipboard
| Challenge: | Open-domain Question Answering (OpenQA) aims at answering factual questions using an external large-scale knowledge corpus. |
| Approach: | They propose a retrieval-augmented approach to QA that focuses on retrieving relevant knowledge from an external corpus. |
| Outcome: | The proposed model can generalize to completely different knowledge domains while adapting to updated versions of the same knowledge corpus and switching to completely new knowledge domain. |
OASIS: Online Sample Selection for Continual Instruction Tuning (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods for continual instruction tuning (CIT) use pre-trained reference models, which are impractical in CIT setups since future data are unknown. |
| Approach: | They propose an adaptive online sample selection approach that estimates each sample’s informativeness relative to all previously seen data and minimizes informative redundancy through iterative selection score updates. |
| Outcome: | Experiments on various large foundation models show that using only 25% of the data achieves comparable performance to full-data training and outperforms the state-of-the-art sampling methods. |
Distill Visual Chart Reasoning Ability from LLMs to MLLMs (2025.findings-emnlp)
Copied to clipboard
Wei He, Zhiheng Xi, Wanxu Zhao, Xiaoran Fan, Yiwen Ding, Zifei Shan, Tao Gui, Qi Zhang, Xuanjing Huang
| Challenge: | a new method for generating chart annotations is proposed to improve visual reasoning in multimodal large language models. |
| Approach: | They propose a code-as-intermediary translation method for distilling visual reasoning abilities from LLMs to MLLMs. |
| Outcome: | The proposed method is cost-effective, efficient and scalable. |
CITB: A Benchmark for Continual Instruction Tuning (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods for instruction tuning do not leverage the rich natural language instructions. |
| Approach: | They propose to use a benchmark to study how instruction tuning works in CL tasks. |
| Outcome: | The proposed method can achieve similar or better results than existing CL methods. |
Don’t Half-listen: Capturing Key-part Information in Continual Instruction Tuning (2025.acl-long)
Copied to clipboard
Yongquan He, Wenyuan Zhang, Xuancheng Huang, Peng Zhang, Lingxun Meng, Xiang Zhou, Ke Zeng, Xunliang Cai
| Challenge: | Existing methods to improve instruction tuning for large language models may cause catastrophic forgetting (CF) CF is a problem where previously learned abilities are degraded . |
| Approach: | They propose a continual instruction tuning method that uses key-part information gain to replay data and refine training objective. |
| Outcome: | The proposed method achieves superior performance on both seen and held-out tasks. |
Advancing SMoE for Continuous Domain Adaptation of MLLMs: Adaptive Router and Domain-Specific Loss (2025.acl-long)
Copied to clipboard
| Challenge: | Recent studies have explored Continual Instruction Tuning (CIT) in Multimodal Large Language Models (MLLMs), with a primary focus on Task-incremental CIT, where MLLM are required to continuously acquire new tasks. |
| Approach: | They propose a Sparse Mixture of Expert (SMoE) based method for domain-incremental CIT in Multimodal Large Language Models (MLLMs) . they equip the SMoA module with a domain-specific autoregressive loss (DSAL) they establish a new benchmark to evaluate the efficacy of their method . |
| Outcome: | The proposed method outperforms all baselines and is based on a Sparse Mixture of Experts (SMoE) module . |
PASs-MoE: Mitigating Misaligned Co-drift among Router and Experts via Pathway Activation Subspaces for Continual Learning (2026.acl-long)
Copied to clipboard
| Challenge: | Existing LoRA-based Mixture-of-Experts (MoE) methods often jointly update the router and experts in an indiscriminate way, causing the router’s preferences to co-drift with experts’ adaptation pathways and exacerbate forgetting. |
| Approach: | They propose a LoRA-induced subspace that reflects which low-rank pathway directions an input activates in each expert, providing a capability-aligned coordinate system for routing and preservation. |
| Outcome: | The proposed method outperforms conventional continual learning baselines and MoE–LoRA variants in accuracy and resistance to forgetting, without increasing model parameters. |