Fast Context Adaptation via Meta-Learning

By Luisa M. Zintgraf, Kyriacos Shiarlis, Vitaly Kurin, Katja Hofmann, Shimon Whiteson. 2018

432 citations recorded by the source corpus.

Subfields: Optimization & Training, 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 Optimization & Training

In Meta & Continual Learning

Citations in the corpus

Flamingo: a Visual Language Model for Few-Shot Learning (2022, methodology): “In this paper, we take inspiration from the in-context [11] few-shot learning technique instead of more involved few-shot learning approaches based on metric learning [24, 103, 112, 117] or meta-learning [6, 7, 27, 31, 91, 155].”
ORBIT: A Real-World Few-Shot Dataset for Teachable Object Recognition (2021, background): “In optimization-based approaches, the model takes many [51, 46, 5] or few [9, 52, 2] gradient steps on the context examples, and the updated model then classifies the target examples.”
Personalized Federated Learning: A Meta-Learning Approach (2020, background): “…was ﬁrst proposed by Finn et al. (2017) and followed by a number of papers studying its empirical characteristics (Antoniou et al., 2019; Li et al., 2017; Grant et al., 2018; Nichol et al., 2018; Zintgraf et al., 2019; Behl et al., 2019) as well as its convergence properties (Fallah et al., 2019).”
CoMPS: Continual Meta Policy Search (2021, background): “, 2019; Dorfman & Tamar, 2020), applying gradient descent from initial parameters (Finn et al., 2017; Gupta et al., 2018b; Rothfuss et al., 2018; Clavera et al., 2018; Mendonca et al., 2019; Zintgraf et al., 2019; Mitchell et al., 2020), and training a critic to provide policy gradients (Sung et al.”
Few-Shot and Continual Learning with Attentive Independent Mechanisms (2021, methodology): “To demonstrate the potential of AIM as a fundamental building block for fast learning without forgetting, we demonstrate its strength on few-shot classification [12, 43, 53] and continual learning [5, 22, 23] benchmarks.”
Continuous Coordination As a Realistic Scenario for Lifelong Learning (2021, methodology): “Moreover, exploiting recent advances in MetaRL such as (Zintgraf et al., 2019) for faster adaptation in the few-shot evaluation setting, instead of naively fine-tuning can lead to agents that adapt well in the ad-hoc scenario.”
Towards Continual Reinforcement Learning: A Review and Perspectives (2020, background): “Current work also typically assumes a stationary task state distribution for meta-learning (Doshi-Velez and Konidaris, 2013; Wang et al., 2016; Duan et al., 2016; Zintgraf et al., 2018; Rakelly et al., 2019; Zintgraf et al., 2019; Humplik et al., 2019; Fakoor et al., 2019; Ortega et al., 2019; Perez”
Embracing Change: Continual Learning in Deep Neural Networks. (2020, neutral): “In [86,87] both loops use gradient descent , while inner-loops can also be nearest-neighbor matching [88,89], recurrent neural representations [8,90,91], probabilistic inference [82,92], or mixed approaches [93 – 96].”
Meta-Consolidation for Continual Learning (2020, methodology): “Meta-learning algorithms - which mostly have focused on the standard learning setting (not continual learning) - can be broadly classified into black-box adaptation based methods [70, 55], optimization based [23, 57, 90] and non-parametric methods [83, 75, 77].”
Task Agnostic Continual Learning via Meta Learning (2019, methodology): “In Fast Context Adaptation via Meta-Learning (CAVIA) (Zintgraf et al., 2019), a neural network model ˆ f takes a context vector c t as an additional input: ˆ y = ˆ f ( x, c t ; θ ) .”
Rapid Learning or Feature Reuse? Towards Understanding the Effectiveness of MAML (2019, methodology): “Prior works (Zintgraf et al., 2018; Javed and White, 2019) have proposed algorithms where some parameters are only updated in the outer loop and others only in the inner loop.”
Meta-Learning in Neural Networks: A Survey (2020, background): “Classification The most common application of metalearning thus far is few-shot multi-class classification in image recognition, where the inner and outer loss functions are typically the cross entropy over training and validation data respectively [19], [20], [41], [73], [75], [76], [86], [88], [97”
Meta-Learning with Warped Gradient Descent (2019, methodology): “Similarly, CAVIA (Zintgraf et al., 2019) meta-learn ﬁxed parameters to facilitate adaptation of a small set of task parameters.”
Deep Reinforcement Learning amidst Continual Structured Non-Stationarity (2021, methodology): “Within meta-reinforcement learning, two dominant techniques exist: optimization-based (Finn et al., 2017; Rothfuss et al., 2019; Zintgraf et al., 2019; Stadie et al., 2018) and context-based, which includes both recurrent architectures (Duan et al., 2016; Wang et al., 2016; Mishra et al., 2018) and…”
Skill-based Meta-Reinforcement Learning (2022, background): “…et al., 2018; Vuorio et al., 2018; Nagabandi et al., 2019; Clavera et al., 2019; 2018; Rakelly et al., 2019; Vuorio et al., 2019; Yang et al., 2019; Zintgraf et al., 2019; Humplik et al., 2019; Zintgraf et al., 2020; Liu et al., 2021) hold the promise of allowing learning agents to quickly adapt…”
Disability-first Dataset Creation: Lessons from Constructing a Dataset for Teachable Object Recognition with Blind and Low Vision Data Collectors (2021, background): “Achieving this vision, however, requires innovation in few-shot ML techniques [13, 18, 55, 61] that are optimised to work on small amounts of real-world data from people who are blind and low vision.”
Meta-Learning with Implicit Gradients (2019, background): “and produce weight updates [25, 5, 33, 48] or predictions for new inputs [50, 12, 58, 40, 38], and optimization-based approaches that use bi-level optimization to embed learning procedures, such as gradient descent, into the meta-optimization problem [15, 13, 8, 60, 34, 17, 59, 23].”
Learning a Universal Template for Few-shot Dataset Generalization (2021, background): “A particular variant of the influential Model Agnostic Meta Learning (MAML) model that is especially related to FLUTE is CAVIA (Zintgraf et al., 2019) that also divides the feature extractor’s parameters into task-general and taskspecific ones.”
Learning Compositional Neural Programs with Recursive Tree Search and Planning (2019, background): “Recent manifestations of this idea of using an embedding to re-program a core neural network to facilitate meta-learning include Zintgraf et al. [2019] and Chen et al. [2019].”
Bootstrapped Meta-Learning (2021, background): “The meta-objective is the validation loss in expectation over the task distribution: Several works have extended this setup by altering the update rule ( ϕ ) (Lee & Choi, 2018; Zintgraf et al., 2019; Park & Oliva, 2019; Flennerhag et al., 2020).”