Challenge: Using the counterfactual memorisation metric, we find that when training neural networks, models will memorise some inputs but not others.
Approach: They use the counterfactual memorisation metric to build a resource that places 5M NMT datapoints on a memorisations-generalisation map and describe how the datapoint’s surface-level characteristics and a models’ per-datum training signals are predictive of memorising in NMT.
Outcome: The proposed model places 5M NMT datapoints on a memorisation-generalisation map and shows how their surface-level characteristics and models’ per-datum training signals are predictive of memorising in NMT.

Similar Papers

Generalisation First, Memorisation Second? Memorisation Localisation for Natural Language Classification Tasks (2024.findings-acl)

Copied to clipboard

Challenge: Memorisation in neural models is concerned due to overfitting and privacy concerns . a dominant hypothesis based on image classification is that lower layers learn generalisable features and deeper layers specialise and memorise.
Approach: They apply 4 techniques to localise and edit models' memories.
Outcome: The proposed method shows that memorisation is a gradual process rather than a localised one.
On Compositional Generalization of Neural Machine Translation (2021.acl-long)

Copied to clipboard

Challenge: Modern neural machine translation models have shown competitive performance in benchmarks such as WMT, but there are significant issues such as robustness, domain generalization, etc.
Approach: They propose a benchmark dataset for NMT models from the perspective of compositional generalization and quantitatively analyze the results.
Outcome: The proposed model performs well under traditional metrics, but is low in out-of-domain and low-resource conditions.
Domain adapted machine translation: What does catastrophic forgetting forget and why? (2024.emnlp-main)

Copied to clipboard

Challenge: Neural Machine Translation (NMT) models can be specialized by domain adaptation, often fine-tuning on a dataset of interest.
Approach: They propose a novel approach to understanding catastrophic forgetting during NMT adaptation by investigating the relationship between the data and the in-domain vocabulary coverage.
Outcome: The proposed model can be specialized by fine-tuning on a domain of interest, but can fail to achieve the predicted quality of the target domain.
Neuron-Level Differentiation of Memorization and Generalization in Large Language Models (2025.emnlp-main)

Copied to clipboard

Challenge: Existing models exhibit memorization and generalization behaviors in ways that are not easily interpretable or controllable.
Approach: They propose to use a GPT-2 and LLaMA-3.2 model to identify distinct neuron subsets responsible for each behavior to steer the model toward memorization or generalization.
Outcome: The proposed models show that inference-time interventions on these neurons can steer the model’s behavior toward memorization or generalization.
Is MAP Decoding All You Need? The Inadequacy of the Mode in Neural Machine Translation (2020.coling-main)

Copied to clipboard

Challenge: Recent studies have revealed a number of pathologies of neural machine translation systems.
Approach: They propose to use maximum a posteriori decoding to identify the highest-scoring translation, i.e. the mode problem, to validate the model and its training algorithm.
Outcome: The proposed model reproduces the statistical data well, but the beam search strays from the statistics.
Finding Memo: Extractive Memorization in Constrained Sequence Generation Tasks (2022.findings-emnlp)

Copied to clipboard

Challenge: Memorization presents a challenge for constrained Natural Language Generation tasks . previous studies focused on counterfactual memorization, linking it to hallucinations .
Approach: They propose an algorithm for extractive memorization in constrained sequence generation tasks . they propose to elicit non-memorized translations of memorized samples from the same model .
Outcome: The proposed algorithm could be leveraged to mitigate memorization in the model through finetuning.
Memorisation versus Generalisation in Pre-trained Language Models (2022.acl-long)

Copied to clipboard

Challenge: State-of-the-art pre-trained language models have been shown to memorise facts and perform well with limited amounts of training data.
Approach: They propose to extend pre-trained language models to generalise and memorise facts in noisy and low-resource scenarios.
Outcome: The proposed extension improves performance in low-resource named entity recognition tasks.
Beyond Noise: Mitigating the Impact of Fine-grained Semantic Divergences on Neural Machine Translation (2021.acl-long)

Copied to clipboard

Challenge: Prior work treats all types of mismatches between source and target as noise . Consequently, it remains unclear how noisy parallel training samples impact NMT training.
Approach: They propose a divergent-aware NMT framework that uses factors to help NMT recover from the degradation caused by naturally occurring divergences.
Outcome: The proposed framework improves translation quality and model calibration on EN-FR tasks.
Investigating Catastrophic Forgetting During Continual Training for Neural Machine Translation (2020.coling-main)

Copied to clipboard

Challenge: Neural machine translation models suffer from catastrophic forgetting during continual training . models tend to overfit to frequent observations in the in-domain data but forget previously learned knowledge.
Approach: They investigated the causes of catastrophic forgetting in NMT models by examining their parameters and modules.
Outcome: The proposed model forgets previously learned knowledge and swings to fit new data . the results show that some parameters are important for both the general-domain and in-domain translation and the great change of them during continual training brings about the performance decline in general- domain.
Categorizing Semantic Representations for Neural Machine Translation (2022.coling-1)

Copied to clipboard

Challenge: Modern neural machine translation models suffer limitation in compositional generalization, resulting in weakened translation performance on unseen compounds.
Approach: They propose to introduce categorization to the contextualized representations to improve generalization by reducing sparsity and overfitting.
Outcome: The proposed method reduces compositional generalization error rates by 24% on a dedicated MT dataset.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations