United States

Incrementally fine-tuning foundational models on new tasks or domains is now the de facto approach in NLP. A known pitfall of this approach is the *catastrophic forgetting* of prior knowledge that happens during fine-tuning. A common approach to alleviate such forgetting is to rehearse samples from prior tasks during fine-tuning. Several existing works assume a fixed memory buffer to store prior task examples, while relying on inferences (forward passes) with the model at hand for choosing examples for rehearsal from the buffer. However, given the increasing computational cost of model inference, and decreasing cost of data storage, we focus on the setting to rehearse samples with a fixed computational budget instead of a fixed memory budget. We propose a sampling scheme, `**mix-cd`, that prioritizes rehearsal of ``collateral damage&#39;&#39; samples, which are samples predicted correctly by the prior model but forgotten by the incrementally tuned one. The crux of our scheme is a procedure to efficiently estimate the density of collateral damage samples without incurring additional model inferences. Our approach is computationally efficient, easy to implement, and outperforms several leading continual learning methods in compute-constrained settings. All the code will be publicly available at https://github.com/jybai/mix-cd-rehearsal.

NAACL 2025

An Efficient Rehearsal Scheme for Catastrophic Forgetting Mitigation during Multi-stage Fine-tuning

online rehearsal

sequential training

catastrophic forgetting

### Welcome!
The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. Our Virtual Poster Sessions will take place online Thursday, August 22, 2024.

You are required to register for this event. **Please register [here](https://2024.aclweb.org/registration). **

If you have already registered, please check your inbox for an email from Underline granting you access to ACL 2024 content.

Please register!

ACL 2024

The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. More information will be announced soon.

Incrementally fine-tuning foundational models on new tasks or domains is now the de facto approach in NLP. A known pitfall of this approach is the *catastrophic forgetting* of prior knowledge that happens during fine-tuning. A common approach to alleviate such forgetting is to rehearse samples from prior tasks during fine-tuning. Several existing works assume a fixed memory buffer to store prior task examples, while relying on inferences (forward passes) with the model at hand for choosing examples for rehearsal from the buffer. However, given the increasing computational cost of model inference, and decreasing cost of data storage, we focus on the setting to rehearse samples with a fixed computational budget instead of a fixed memory budget. We propose a sampling scheme, `**mix-cd`, that prioritizes rehearsal of ``collateral damage'' samples, which are samples predicted correctly by the prior model but forgotten by the incrementally tuned one. The crux of our scheme is a procedure to efficiently estimate the density of collateral damage samples without incurring additional model inferences. Our approach is computationally efficient, easy to implement, and outperforms several leading continual learning methods in compute-constrained settings. All the code will be publicly available at https://github.com/jybai/mix-cd-rehearsal.

poster

### Welcome to 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics

Welcome to the 2025 meeting of the Nations of the Americas Chapter of the Association for Computational Linguistics! I am proud to help organize the first NAACL conference to carry the new name of our organization, one that emphasizes inclusion for all of the Americas. I am also pleased to welcome you to Albuquerque, New Mexico, a state whose unique blend of cultural influences will make for an excellent backdrop for NAACL 2025, especially with this year’s special theme on NLP in a Multicultural World. 
**[Continue reading...](https://drive.google.com/file/d/1jX-qGhqVSZZCIrAnJaz798pflu5Irrdn/view?usp=sharing)**

*- Colin Cherry, Google, NAACL 2025 General Chair* 

[![](https://assets.underline.io/markdown_image/1/image/b087f8a4dc5816d6e1a6514e59c59ac3.png)](https://drive.google.com/file/d/1T96GzPqObXrMln2BMByCSXTSizjTg69P/view?usp=sharing)

You need to log in with the email address you registered with. Access credentials have been sent to your email. 

Please be sure to check your spam and other email folders if you do not see an email confirmation right away.

Please log in to explore this event.

To access NAACL 2025 event page you are required to register. Please follow [**this link**](https://2025.naacl.org/registration/registration/) to register. Access will depend on your registration type.

Welcome to 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics

Large language models (LLMs) have shown remarkable performance on code generation tasks. A recent use case is iterative code repair, where an LLM fixes an incorrect program by rationalizing about errors and generating new code. Recent works augment the code repair process by integrating modern techniques such as chain-of-thought reasoning or distillation, but only study their benefits on high-resource languages like Python, and ignore low-resource languages like Perl. To address this gap of knowledge, we investigate the benefits of distilling code repair for both high and low resource languages to determine if the techniques that are effective in a high resource setting are also applicable in a low resource setting. Our evaluation shows that distilling the ability to repair code has language dependent benefits. To explain this behavior, we perform a further analysis and find that contrary to preexisting beliefs, the correlation between reasoning ability and code correction ability is weak. We hypothesize this weak correlation is magnified in low-resource settings where base models lack deep knowledge of a programming language, leading to wavering benefits of code repair.

Investigating the Transferability of Code Repair for Low-Resource Programming Languages

Despite the success of Large Language Models (LLMs), they still face challenges related to high inference costs and memory requirements. To address these issues, Knowledge Distillation (KD) has emerged as a popular method for model compression, with the use of student-generated outputs (SGOs) as training data being particularly notable for reducing the mismatch between training and inference. However, SGOs often produce noisy and biased sequences, which can lead to misguidance from the teacher model, especially in long sequences. To mitigate these challenges, we propose SWITCH (Studying With Teacher for Knowledge Distillation), a novel approach that strategically incorporates the teacher model during the student’s sequence generation. SWITCH identifies discrepancies between the token probabilities of the teacher and student models, allowing the teacher to intervene selectively, particularly in long sequences that are more prone to teacher misguidance. Extensive experimental results across three model families and five instruction-following datasets show that SWITCH surpasses traditional KD methods, particularly excelling in the generation of long sequential data.

SWITCH: Studying with Teacher for Knowledge Distillation of Large Language Models

To efficiently select optimal dataset combinations for enhancing multi-task learning (MTL) performance in large language models, we proposed a novel framework that leverages a neural network to predict the best dataset combinations. The framework iteratively refines the selection, greatly improving efficiency, while being model-, dataset-, and domain-independent. Through experiments on 12 biomedical datasets across four tasks—named entity recognition, relation extraction, event extraction, and text classification—we demonstrate that our approach effectively identifies better combinations, even for tasks that may seem unpromising from a human perspective. This verifies that our framework provides a promising solution for maximizing MTL potential.

Towards Better Multi-task Learning: A Framework for Optimizing Dataset Combinations in Large Language Models

Large Language Models (LLMs) excel in many Natural Language Processing tasks but are outperformed by specialized tools for certain tasks. This raises the question: Can we reduce redundant LLM parameters when using these tools? Given the size and high training costs of LLMs, it is essential to efficiently forget specific knowledge without retraining. This paper introduces UNLEARN, a novel method that uses subspace techniques to selectively remove knowledge without access to the original training data, without retraining, and with minimal impact to other tasks. Our results show that UNLEARN significantly outperforms previous methods for forgetting targeted (unwanted) knowledge while also preserving related (wanted) knowledge. We also propose LEARN, a complementary approach for targeted knowledge addition, which achieves fine-tuning accuracy comparable to Low-Rank Adaptation (LoRA) without degrading related task performance.

UNLEARN Efficient Removal of Knowledge in Large Language Models

Over the last years, state-of-the-art AI models have grown to a point where their use bears significant economic and environmental cost. At the same time, investigation of NLP models has shown that they are often overparameterized, giving rise to research of compression approaches. Such approaches often suffer the trade-off between hardware requirements and classification performance. In this work, we propose the hardware-independent compression strategy Adaptive Parameter Compression (APC). We extend the Weight Squeezing approach by introducing compression biases and weights, as well as investigating multiple initialization strategies for these weights and the application of APC to transformer model components. Experiments with BERT_base show the compression's effectiveness, slightly outperforming DistilBERT while being significantly more efficient.

Adaptive Parameter Compression for Language Models

Causal inference has demonstrated significant potential to enhance Natural Language Processing (NLP) models in areas such as predictive accuracy, fairness, robustness, and explainability by capturing causal relationships among variables. The rise of generative Large Language Models (LLMs) has greatly impacted various language processing tasks. This survey focuses on research that evaluates or improves LLMs from a causal view in the following areas: reasoning capacity, fairness and safety issues, explainability, and handling multimodality. Meanwhile, LLMs can assist in causal inference tasks, such as causal relationship discovery and causal effect estimation, by leveraging their generation ability and knowledge learned during pre-training. This review explores the interplay between causal inference frameworks and LLMs from both perspectives, emphasizing their collective potential to further the development of more advanced and robust artificial intelligence systems.

Large Language Models and Causal Inference in Collaboration: A Comprehensive Survey

While Large Language Models (LLMs) exhibit remarkable capabilities in zero-shot and few-shot scenarios, they often require computationally prohibitive sizes. Conversely, smaller Masked Language Models (MLMs) like BERT and RoBERTa achieve state-of-the-art results through fine-tuning but struggle with extending to few-shot and zero-shot settings due to their architectural constraints. Hence, we propose Statement-Tuning, a technique that models discriminative tasks as a set of finite statements and trains an encoder model to discriminate between the potential statements to determine the label. We do Statement-Tuning on multiple tasks to enable cross-task generalization. Experimental results demonstrate that Statement-Tuning achieves competitive performance compared to state-of-the-art LLMs with significantly fewer parameters. Furthermore, we compare with previous encoder-based methodology and show that our method is more accurate and more robust to spurious patterns. Moreover, the study investigates the impact of several design choices on few-shot and zero-shot generalization, revealing that Statement-Tuning can achieve strong performance with modest training data and benefits from task and statement diversity for unseen task generalizability. We release all the code used to generate statement data, train and evaluate our Statement-Tuned models.

Enabling Natural Zero-Shot Prompting on Encoder Models via Statement-Tuning

Large language models (LLMs) have demonstrated significant progress in various natural language generation and understanding tasks. However, their linguistic generalization capabilities remain questionable, raising doubts about whether these models learn language similarly to humans. While humans exhibit compositional generalization and linguistic creativity in language use, the extent to which LLMs replicate these abilities, particularly in morphology, is under-explored. In this work, we systematically investigate the morphological generalization abilities of LLMs through the lens of compositionality. We define morphemes as compositional primitives and design a novel suite of generative and discriminative tasks to assess morphological productivity and systematicity. Focusing on agglutinative languages such as Turkish and Finnish, we evaluate several state-of-the-art instruction-finetuned multilingual models, including GPT-4 and Gemini. Our analysis shows that LLMs struggle with morphological compositional generalization particularly when applied to novel word roots, with performance declining sharply as morphological complexity increases. While models can identify individual morphological combinations better than chance, their performance lacks systematicity, leading to significant accuracy gaps compared to humans.

Evaluating Morphological Compositional Generalization in Large Language Models

Decomposition of text into atomic propositions is a flexible framework allowing for the closer inspection of input and output text. We use atomic decomposition of hypotheses in two natural language reasoning tasks, traditional NLI and defeasible NLI, to form atomic sub-problems, or granular inferences that models must weigh when solving the overall problem. These atomic sub-problems serve as a tool to further understand the structure of both NLI and defeasible reasoning, probe a model's consistency and understanding of different inferences, and measure the diversity of examples in benchmark datasets. Our results indicate that LLMs still struggle with logical consistency on atomic NLI and defeasible NLI sub-problems. Lastly, we identify critical atomic sub-problems of defeasible NLI examples, or those that most contribute to the overall label, and propose a method to measure the inferential consistency of a model, a metric designed to capture the degree to which a model makes consistently correct or incorrect predictions about the same fact under different contexts.

NLI under the Microscope: What Atomic Hypothesis Decomposition Reveals

Allophony refers to the variation in the phonetic realization of a phoneme based on its phonetic environment. Modeling allophones is crucial for atypical pronunciation assessment, which involves distinguishing atypical from typical pronunciations. However, recent phoneme classifier-based approaches often simplify this by treating various realizations as a single phoneme, bypassing the complexity of modeling allophonic variation. Motivated by the acoustic modeling capabilities of frozen self-supervised speech model (S3M) features, we propose MixGoP, a novel approach that leverages Gaussian mixture models to model phoneme distributions with multiple subclusters. Our experiments show that MixGoP achieves state-of-the-art performance across four out of five datasets, including dysarthric and non-native speech. Our analysis further suggests that S3M features capture allophonic variation more effectively than MFCCs and Mel spectrograms, highlighting the benefits of integrating MixGoP with S3M features.

Downloads

Next from NAACL 2025

Investigating the Transferability of Code Repair for Low-Resource Programming Languages

Similar lecture

Mitigating Catastrophic Forgetting in Large Language Models with Self-Synthesized Rehearsal