# IIRC: Incremental Implicitly-Refined Classification

arXiv: https://arxiv.org/abs/2012.12477 · HTML: https://mlsanity.com/paper/2012.12477/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Dec 2020.

## Built on by (most cited first, 20 of 20)

- [A Comprehensive Survey of Continual Learning: Theory, Method and Application](https://mlsanity.com/paper/2302.00487/) (2023)
- [Class-Incremental Learning: A Survey](https://mlsanity.com/paper/2302.03648/) (2023)
- [An Empirical Investigation of the Role of Pre-training in Lifelong Learning](https://mlsanity.com/paper/2112.09153/) (2021)
- [Recent Advances of Continual Learning in Computer Vision: An Overview](https://mlsanity.com/paper/2109.11369/) (2021)
- [Understanding Continual Learning Settings with Data Distribution Drift Analysis](https://mlsanity.com/paper/2104.01678/) (2021)
- [Long-Tailed Class Incremental Learning](https://mlsanity.com/paper/2210.00266/) (2022)
- [Continuum: Simple Management of Complex Continual Learning Scenarios](https://mlsanity.com/paper/2102.06253/) (2021)
- [Continuous Coordination As a Realistic Scenario for Lifelong Learning](https://mlsanity.com/paper/2103.03216/) (2021)
- [Multi-label Iterated Learning for Image Classification with Label Ambiguity](https://mlsanity.com/paper/2111.12172/) (2021)
- [Continual Learning with Evolving Class Ontologies](https://mlsanity.com/paper/2210.04993/) (2022)
- [Beyond Supervised Continual Learning: a Review](https://mlsanity.com/paper/2208.14307/) (2022)
- [Continual Adapter Tuning with Semantic Shift Compensation for Class-Incremental Learning](https://mlsanity.com/paper/2403.19979/) (2024)
- [Coarse-To-Fine Incremental Few-Shot Learning](https://mlsanity.com/paper/2111.14806/) (2021)
- [DioR: Adaptive Cognitive Detection and Contextual Retrieval Optimization for Dynamic Retrieval-Augmented Generation](https://mlsanity.com/paper/2504.10198/) (2025)
- [Multi-Teacher Knowledge Distillation for Incremental Implicitly-Refined Classification](https://mlsanity.com/paper/2202.11384/) (2022)
- [HCV: Hierarchy-Consistency Verification for Incremental Implicitly-Refined Classification](https://mlsanity.com/paper/2110.11148/) (2021)
- [Fast Hierarchical Learning for Few-Shot Object Detection](https://mlsanity.com/paper/2210.05008/) (2022)
- [Towards Continually Learning Application Performance Models](https://mlsanity.com/paper/2310.16996/) (2023)
- [An Attention-based Representation Distillation Baseline for Multi-Label Continual Learning](https://mlsanity.com/paper/2407.14249/) (2024)
- [Online Continual Learning with Dynamic Label Hierarchies](https://mlsanity.com/paper/2605.11742/) (2026)

## Abstract (arXiv)

We introduce the "Incremental Implicitly-Refined Classi-fication (IIRC)" setup, an extension to the class incremental learning setup where the incoming batches of classes have two granularity levels. i.e., each sample could have a high-level (coarse) label like "bear" and a low-level (fine) label like "polar bear". Only one label is provided at a time, and the model has to figure out the other label if it has already learnfed it. This setup is more aligned with real-life scenarios, where a learner usually interacts with the same family of entities multiple times, discovers more granularity about them, while still trying not to forget previous knowledge. Moreover, this setup enables evaluating models for some important lifelong learning challenges that cannot be easily addressed under the existing setups. These challenges can be motivated by the example "if a model was trained on the class bear in one task and on polar bear in another task, will it forget the concept of bear, will it rightfully infer that a polar bear is still a bear? and will it wrongfully associate the label of polar bear to other breeds of bear?". We develop a standardized benchmark that enables evaluating models on the IIRC setup. We evaluate several state-of-the-art lifelong learning algorithms and highlight their strengths and limitations. For example, distillation-based methods perform relatively well but are prone to incorrectly predicting too many labels per image. We hope that the proposed setup, along with the benchmark, would provide a meaningful problem setting to the practitioners