# On the Impact of Temporal Concept Drift on Model Explanations

arXiv: https://arxiv.org/abs/2210.09197 · HTML: https://mlsanity.com/paper/2210.09197/

## Key facts

- Cited by 14 papers in the map (data through 14 Oct 2026).
- Builds on 37 of its 69 references that are in the map.
- First posted on arXiv 17 Oct 2022.

## Builds on (most cited first, 30 of 37)

- [A Unified Approach to Interpreting Model Predictions](https://mlsanity.com/paper/1705.07874/) (2017)
- ["Why Should I Trust You?": Explaining the Predictions of Any Classifier](https://mlsanity.com/paper/1602.04938/) (2016)
- [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017)
- [Character-level Convolutional Networks for Text Classification](https://mlsanity.com/paper/1509.01626/) (2015)
- [Learning Important Features Through Propagating Activation Differences](https://mlsanity.com/paper/1704.02685/) (2017)
- [A Survey Of Methods For Explaining Black Box Models](https://mlsanity.com/paper/1802.01933/) (2018)
- [Attention is not Explanation](https://mlsanity.com/paper/1902.10186/) (2019)
- [Attention is not not Explanation](https://mlsanity.com/paper/1908.04626/) (2019)
- [Towards Faithfully Interpretable NLP Systems: How should we define and evaluate faithfulness?](https://mlsanity.com/paper/2004.03685/) (2020)
- [ERASER: A Benchmark to Evaluate Rationalized NLP Models](https://mlsanity.com/paper/1911.03429/) (2019)
- [A Benchmark for Interpretability Methods in Deep Neural Networks](https://mlsanity.com/paper/1806.10758/) (2018)
- [Is Attention Interpretable?](https://mlsanity.com/paper/1906.03731/) (2019)
- [Calibration of Pre-trained Transformers](https://mlsanity.com/paper/2003.07892/) (2020)
- [Time-Aware Language Models as Temporal Knowledge Bases](https://mlsanity.com/paper/2106.15110/) (2021)
- [Rationalizing Neural Predictions](https://mlsanity.com/paper/1606.04155/) (2016)
- [Mind the Gap: Assessing Temporal Generalization in Neural Language Models](https://mlsanity.com/paper/2102.01951/) (2021)
- [A Diagnostic Study of Explainability Techniques for Text Classification](https://mlsanity.com/paper/2009.13295/) (2020)
- [TimeLMs: Diachronic Language Models from Twitter](https://mlsanity.com/paper/2202.03829/) (2022)
- [Interpretable Neural Predictions with Differentiable Binary Variables](https://mlsanity.com/paper/1905.08160/) (2019)
- [Learning to Faithfully Rationalize by Construction](https://mlsanity.com/paper/2005.00115/) (2020)
- [Analysing Lexical Semantic Change with Contextualised Word Representations](https://mlsanity.com/paper/2004.14118/) (2020)
- [X-FACT: A New Benchmark Dataset for Multilingual Fact Checking](https://mlsanity.com/paper/2106.09248/) (2021)
- [We Need to Talk About Random Splits](https://mlsanity.com/paper/2005.00636/) (2020)
- [Temporal Analysis of Language through Neural Language Models](https://mlsanity.com/paper/1405.3515/) (2014)
- [Investigating the influence of noise and distractors on the interpretation of neural networks](https://mlsanity.com/paper/1611.07270/) (2016)
- [Temporal Adaptation of BERT and Performance on Downstream Document Classification: Insights from Social Media](https://mlsanity.com/paper/2104.08116/) (2021)
- [Temporal Effects on Pre-trained Models for Language Processing Tasks](https://mlsanity.com/paper/2111.12790/) (2021)
- [Evaluating and Characterizing Human Rationales](https://mlsanity.com/paper/2010.04736/) (2020)
- [Evaluating Recurrent Neural Network Explanations](https://mlsanity.com/paper/1904.11829/) (2019)
- [Improving the Faithfulness of Attention-based Explanations with Task-specific Information for Text Classification](https://mlsanity.com/paper/2105.02657/) (2021)

## Built on by (most cited first, 14 of 14)

- [Dynamic Benchmarking of Masked Language Models on Temporal Concept Drift with Multiple Views](https://mlsanity.com/paper/2302.12297/) (2023)
- [NEO-BENCH: Evaluating Robustness of Large Language Models with Neologisms](https://mlsanity.com/paper/2402.12261/) (2024)
- [Investigating Hallucinations in Pruned Large Language Models for Abstractive Summarization](https://mlsanity.com/paper/2311.09335/) (2023)
- [It's about Time: Rethinking Evaluation on Rumor Detection Benchmarks using Chronological Splits](https://mlsanity.com/paper/2302.03147/) (2023)
- [Incorporating Attribution Importance for Improving Faithfulness Metrics](https://mlsanity.com/paper/2305.10496/) (2023)
- [ChronosLex: Time-aware Incremental Training for Temporal Generalization of Legal Classification Tasks](https://mlsanity.com/paper/2405.14211/) (2024)
- [A Survey on Failure Analysis and Fault Injection in AI Systems](https://mlsanity.com/paper/2407.00125/) (2024)
- [PILOT: Legal Case Outcome Prediction with Case Law](https://mlsanity.com/paper/2401.15770/) (2024)
- [Comparing Explanation Faithfulness between Multilingual and Monolingual Fine-tuned Language Models](https://mlsanity.com/paper/2403.12809/) (2024)
- [Fine-Tuned LLMs are "Time Capsules" for Tracking Societal Bias Through Books](https://mlsanity.com/paper/2502.05331/) (2025)
- [Examining Temporal Bias in Abusive Language Detection](https://mlsanity.com/paper/2309.14146/) (2023)
- [Gender Bias in Explainability: Investigating Performance Disparity in Post-hoc Methods](https://mlsanity.com/paper/2505.01198/) (2025)
- [EvalxNLP: A Framework for Benchmarking Post-Hoc Explainability Methods on NLP Models](https://mlsanity.com/paper/2505.01238/) (2025)
- [How often do Answers Change? Estimating Recency Requirements in Question Answering](https://mlsanity.com/paper/2603.16544/) (2026)

## Abstract (arXiv)

Explanation faithfulness of model predictions in natural language processing is typically evaluated on held-out data from the same temporal distribution as the training data (i.e. synchronous settings). While model performance often deteriorates due to temporal variation (i.e. temporal concept drift), it is currently unknown how explanation faithfulness is impacted when the time span of the target data is different from the data used to train the model (i.e. asynchronous settings). For this purpose, we examine the impact of temporal variation on model explanations extracted by eight feature attribution methods and three select-then-predict models across six text classification tasks. Our experiments show that (i)faithfulness is not consistent under temporal variations across feature attribution methods (e.g. it decreases or increases depending on the method), with an attention-based method demonstrating the most robust faithfulness scores across datasets; and (ii) select-then-predict models are mostly robust in asynchronous settings with only small degradation in predictive performance. Finally, feature attribution methods show conflicting behavior when used in FRESH (i.e. a select-and-predict model) and for measuring sufficiency/comprehensiveness (i.e. as post-hoc methods), suggesting that we need more robust metrics to evaluate post-hoc explanation faithfulness.