iTAML: An Incremental Task-Agnostic Meta-learning Approach

By Jathushan Rajasegaran, Salman Hameed Khan, Munawar Hayat, Fahad Shahbaz Khan, Mubarak Shah. 2020

173 citations recorded by the source corpus.

Humans can continuously learn new knowledge as their experience grows. In contrast, previous learning in deep neural networks can quickly fade out when they are trained on a new task. In this paper, we hypothesize this problem can be avoided by learning a set of generalized parameters, that are neither specific to old nor new tasks. In this pursuit, we introduce a novel meta-learning approach that seeks to maintain an equilibrium between all the encountered tasks. This is ensured by a new meta-update rule which avoids catastrophic forgetting. In comparison to previous meta-learning techniques, our approach is task-agnostic. When presented with a continuum of data, our model automatically identifies the task and quickly adapts to it with just a single update. We perform extensive experiments on five datasets in a class-incremental setting, leading to significant improvements over the state of the art methods (e.g., a 21.3% boost on CIFAR100 with 10 incremental tasks). Specifically, on large-scale datasets that generally prove difficult cases for incremental learning, our approach delivers absolute gains as high as 19.1% and 7.4% on ImageNet and MS-Celeb datasets, respectively.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Meta & Continual Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Meta & Continual Learning

Citations in the corpus

A Comprehensive Survey of Continual Learning: Theory, Method and Application (2023, background): “For example, MER [360] aligns their gradient directions, while iTAML [349] applies a metaupdating rule to keep them in balance with each other.”
Learnability and Algorithm for Continual Learning (2023, methodology): “…buffer data (Rusu et al., 2016; LopezPaz & Ranzato, 2017; Chaudhry et al., 2019a; Hou et al., 2019; Wu et al., 2019; Rolnick et al., 2019; Buzzega et al., 2020; Rajasegaran et al., 2020a; Liu et al., 2021; Cha et al., 2021; Yan et al., 2021; Wang et al., 2022b; Guo et al., 2022; Kim et al., 2022a).”
RMM: Reinforced Memory Management for Class-Incremental Learning (2023, background): “…task-incremental learning—each phase corresponds to a task (dataset) that contains new data of all seen classes [7, 10, 19, 29, 43, 47, 57], or class-incremental learning (CIL)—each phase contains data of a new set of classes, i.e., classes are unseen [4, 6, 18, 25, 31, 37, 40, 41, 49, 52, 55–57].”
A Theoretical Study on Solving Continual Learning (2022, methodology): “Some methods have used a TIL method for CIL with an additional task-id prediction technique. iTAML [62] requires each test batch to be from a single task.”
Anti-Retroactive Interference for Lifelong Learning (2022, methodology): “iTAML [21] introduces a meta-learning approach that seeks to maintain an equilibrium between all the encountered tasks, in the sense that it is unbiased towards class samples of majority and simultaneously minimizes forgetting.”
Learning to Imagine: Diversify Memory for Incremental Learning using Unlabeled Data (2022, methodology): “Popular structure-driven methods [2,20,21,25,32,33,35] modify the network structure or expand the network for training new tasks.”
Few-Shot Class-Incremental Learning by Sampling Multi-Phase Tasks (2022, background): “Other works consider saving embeddings instead of raw images [39], using generative models for data rehearsal [40], taskwise adaptation [41], and output normalization [20], [42], [43].”
Overcoming Catastrophic Forgetting in Incremental Few-Shot Learning by Finding Flat Minima (2021, methodology): “Previous research mainly includes multi-class incremental learning [4, 38, 22, 33, 54, 51] and multi-task incremental learning [21, 31, 42].”
Recent Advances of Continual Learning in Computer Vision: An Overview (2021, positive): “[205] proposed an Incremental Task-Agnostic Meta-learning (iTAML) method, which adapts meta learning to separate the task-agnostic feature extractor from the task-specific classifier.”
Generalized and Incremental Few-Shot Learning by Explicit Learning and Calibration without Forgetting (2021, methodology): “To achieve a balanced performance in the joint space of base and novel classes we apply the replay technique that is common in incremental learning [36, 27, 35, 25].”
Optimizing Reusable Knowledge for Continual Learning via Metalearning (2021, methodology): “It has been used as a training strategy in sequence learning [43, 44] and continual learning scenarios [45, 46, 47, 48].”
Few-Shot Class-Incremental Learning via Relation Knowledge Distillation (2021, background): “The scenarios of incremental learning can be brieﬂy divided into the class-incremental (Rebufﬁ et al. 2017; Tao et al. 2020a; Chang et al. 2021) and the task-incremental (Rajasegaran et al. 2020) ones.”
On Learning the Geodesic Path for Incremental Learning (2021, neutral): “[36] propose a meta-learning approach (e.”
Towards Open World Object Detection (2021, methodology): “We note that a number of involved approaches have been developed to alleviate such forgetting, including methods based on parameter regularization [2, 24, 29, 66], exemplar replay [6, 51, 37, 5], dynamically expanding networks [39, 60, 56] and meta-learning [50, 25].”
Online Continual Learning in Image Classification: An Empirical Survey (2021, background): “More recently, iTAML [81] proposes to learn a task-agnostic model that automatically predicts the task and quickly adapts to the predicted task with meta-update.”
Do not Forget to Attend to Uncertainty while Mitigating Catastrophic Forgetting (2021, background): “iTaML [32] proposes a task agnostic meta-learning approach that seeks to maintain an equilibrium between all the encountered tasks.”
Adaptive Aggregation Networks for Class-Incremental Learning (2020, background): “[32] incrementally learned new tasks while learning a generic model to retain the knowledge from all tasks.”
Remembering for the Right Reasons: Explanations Reduce Catastrophic Forgetting (2020, methodology): “This simple method, known as experience replay, has been explored and shown to be effective (Rebuffi et al., 2017; Wu et al., 2019; Castro et al., 2018; Rajasegaran et al., 2020; Ebrahimi et al., 2020b; Hayes et al., 2019; Riemer et al., 2018).”