# Specialization in Hierarchical Learning Systems

@article{Hihn2020SpecializationIH, title={Specialization in Hierarchical Learning Systems}, author={Heinke Hihn and Daniel A. Braun}, journal={Neural Processing Letters}, year={2020}, volume={52}, pages={2319 - 2352} }

Joining multiple decision-makers together is a powerful way to obtain more sophisticated decision-making systems, but requires to address the questions of division of labor and specialization. We investigate in how far information constraints in hierarchies of experts not only provide a principled method for regularization but also to enforce specialization. In particular, we devise an information-theoretically motivated on-line learning rule that allows partitioning of the problem space into…

## 14 Citations

### Hierarchically structured task-agnostic continual learning

- Computer ScienceMachine Learning
- 2022

A task-agnostic view of continual learning is taken and a hierarchical information-theoretic optimality principle is developed that facilitates a trade-off between learning and forgetting and proposes a neural network layer, called the Mixture-of-Variational-Experts layer, that alleviates forgetting.

### Mixture-of-Variational-Experts for Continual Learning

- Computer ScienceArXiv
- 2021

This work takes a task-agnostic view of continual learning and develops a hierarchical information-theoretic optimality principle that facilitates a trade-off between learning and forgetting, and proposes a neural network layer, called the Mixture-of-Variational-Experts layer, that alleviates forgetting.

### Using Meta Labels for the Training of Weighting Models in a Sample-Specific Late Fusion Classification Architecture

- Computer Science2020 25th International Conference on Pattern Recognition (ICPR)
- 2021

A novel late fusion architecture is introduced, which can be interpreted as a combination of the well-known mixture of experts and stacked generalization methods that aggregates the outputs of classification models and corresponding sample-specific weighting models.

### Variational Inference for Model-Free and Model-Based Reinforcement Learning

- Computer ScienceArXiv
- 2022

This manuscript shows how the apparently different subjects of VI and RL are linked in two fundamental ways: first, the optimization objective of RL to maximize future cumulative rewards can be recovered via a VI objective under a soft policy constraint in both the non-sequential and the sequential setting.

### Co-Learning for Few-Shot Learning

- Computer ScienceNeural Processing Letters
- 2022

A Co-learning (CL) method for FSL that tries to exploit two basic classifiers to separately infer pseudo-labels for unlabeled samples, and crossly expand them to the labeled data to make the predicted accuracy more reliable.

### Co-Learning for Few-Shot Learning

- Computer ScienceNeural Process. Lett.
- 2022

A Co-learning (CL) method for FSL that tries to exploit two basic classifiers to separately infer pseudo-labels for unlabeled samples, and crossly expand them to the labeled data to make the predicted accuracy more reliable.

### DMH-FSL: Dual-Modal Hypergraph for Few-Shot Learning

- Computer ScienceNeural Process. Lett.
- 2022

This work introduces hypergraph structure and proposes the Dual-Modal Hypergraph Few-Shot Learning (DMH-FSL) method, which is easy to extend to other graph-based methods and demonstrates the efficiency of the method on three benchmark datasets.

### Binary Classification: Counterbalancing Class Imbalance by Applying Regression Models in Combination with One-Sided Label Shifts

- Computer ScienceICAART
- 2021

A novel method is introduced, which addresses the issues of class imbalance in binary, imbalanced classification tasks, by generating a set of negative and positive target labels, such that the corresponding regression task becomes balanced, with respect to the redefined target label set.

### Predicting the Impact of High-Speed Rail on Population Change in Local Cities by using a Naive Bayesian Classification-based Artificial Intelligence Model

- EconomicsThe Journal of Korean Institute of Information Technology
- 2022

Korea is dealing with a population disparity between the capital areas of Seoul and the population of local cities. As a result, the Korean governments established and implemented regional balancing…

### Multi-Modal Pain Intensity Assessment Based on Physiological Signals: A Deep Learning Perspective

- Computer ScienceFrontiers in Physiology
- 2021

This work introduces several novel multi-modal deep learning approaches (characterized by specific supervised, as well as self-supervised learning techniques) for the assessment of pain intensity based on measurable bio-physiological data.

## References

SHOWING 1-10 OF 102 REFERENCES

### An Information-theoretic On-line Learning Principle for Specialization in Hierarchical Decision-Making Systems

- Computer Science2019 IEEE 58th Conference on Decision and Control (CDC)
- 2019

An information-theoretic principle that drives division of labor and specialization when decision-makers with information constraints are joined together is studied and an on-line learning rule of this principle is devised that learns a partitioning of the problem space such that it can be solved by specialized linear policies.

### Hierarchical Relative Entropy Policy Search

- Computer ScienceAISTATS
- 2012

This work defines the problem of learning sub-policies in continuous state action spaces as finding a hierarchical policy that is composed of a high-level gating policy to select the low-level sub-Policies for execution by the agent and treats them as latent variables which allows for distribution of the update information between the sub- policies.

### Meta-learning of Sequential Strategies

- Computer ScienceArXiv
- 2019

This report recast memory-based meta-learning within a Bayesian framework, showing that the meta-learned strategies are near-optimal because they amortize Bayes-filtered data, where the adaptation is implemented in the memory dynamics as a state-machine of sufficient statistics.

### Multitask Learning

- Computer ScienceMachine Learning
- 2004

Prior work on MTL is reviewed, new evidence that MTL in backprop nets discovers task relatedness without the need of supervisory signals is presented, and new results for MTL with k-nearest neighbor and kernel regression are presented.

### Analyzing Abstraction and Hierarchical Decision-Making in Absolute Identification by Information-Theoretic Bounded Rationality

- PsychologyFront. Neurosci.
- 2019

This work compares subjects' behavior to the maximum efficiency predicated by the bounded rational decision-making model, and finds that subjects adapt their abstraction level depending on the available resources.

### Meta-Learning in Computational Intelligence

- Computer ScienceMeta-Learning in Computational Intelligence
- 2011

This book defines and reveals new theoretical and practical trends in meta-learning, inspiring the readers to further research in this exciting field.

### Systems of Bounded Rational Agents with Information-Theoretic Constraints

- Computer ScienceNeural Computation
- 2019

The results suggest that hierarchical architectures of specialized units at lower levels that are coordinated by units at higher levels are optimal, given that each unit's information-processing capability is limited and conforms to constraints on complexity costs.

### Bounded Rational Decision-Making with Adaptive Neural Network Priors

- Computer ScienceANNPR
- 2018

This work investigates generative neural networks as priors that are optimized concurrently with anytime sample-based decision-making processes such as MCMC, and evaluates this approach on toy examples.

### Bounded Rationality, Abstraction, and Hierarchical Decision-Making: An Information-Theoretic Optimality Principle

- Computer ScienceFront. Robot. AI
- 2015

This work applies the basic principle of this framework of bounded rational decision-making to perception-action systems with multiple information-processing nodes and derive bounded optimal solutions and formalizes a mathematically unifying optimization principle that could potentially be extended to more complex systems.

### Information asymmetry in KL-regularized RL

- Computer ScienceICLR
- 2019

This work starts from the KL regularized expected reward objective and introduces an additional component, a default policy, but crucially restricts the amount of information the default policy receives, forcing it to learn reusable behaviors that help the policy learn faster.