Summary of Towards Cost Sensitive Decision Making, by Yang Li et al.
Towards Cost Sensitive Decision Making
by Yang Li, Junier Oliva
First submitted to arxiv on: 4 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper proposes a novel Reinforcement Learning (RL) framework for decision-making in partially-observed environments, where agents can actively acquire features to improve decision quality and certainty. The proposed Active-Acquisition Partially-Observed Markov Decision Process (AA-POMDP) model enables agents to balance the cost of feature acquisition with the reward of task decision. To alleviate exploration-exploitation dilemmas, the authors develop a model-based approach using deep generative models to impute unobserved features and capture dependencies between features. Hierarchical RL algorithms are then developed to resolve AA-POMDPs. The results demonstrate significant performance improvements over existing POMDP-RL solutions. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper helps computers make better decisions when they don’t have all the information. This is important because in real life, we often need to make choices without knowing everything. Right now, computer models either assume they can get all the information or think of some parts as missing and can’t be found. The authors suggest a new way for these models to actively gather more information from their environment. They call this “Active-Acquisition POMDP”. To help computers make decisions in these situations, the authors propose using deep learning models that can predict what they don’t know. This allows them to balance the cost of gathering more information with the reward of making a good decision. The results show that this new approach works better than existing methods. |
Keywords
* Artificial intelligence * Deep learning * Reinforcement learning
