Research map / Reinforcement learning / Imitation learning
Deeply AggreVaTeD: Differentiable Imitation Learning for Sequential Prediction
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use AggreVaTeD with an LSTM instead of a reactive NN | AggreVaTeD · Handwritten algebra dependency parsing | UAS: 0.851 -> 0.924 | Helped |
| Use AggreVaTeD instead of RL with an LSTM | Handwritten algebra dependency parsing · Handwritten algebra dependency parsing | UAS: 0.256 -> 0.924 | Helped |
| Use AggreVaTeD instead of DAgger | Handwritten algebra dependency parsing · Handwritten algebra dependency parsing | UAS: 0.832±0.02 -> 0.924±0.10 | Helped |
| Use natural gradient instead of regular gradient (ablation) | AggreVaTeD with an LSTM · Handwritten algebra dependency parsing | UAS: 0.924±0.10 -> 0.915±0.10 | Hurt |
| Use natural gradient instead of regular gradient (ablation) | AggreVaTeD with an NN · Handwritten algebra dependency parsing | UAS: 0.851±0.10 -> 0.800±0.10 | Hurt |
| Use AggreVaTeD instead of RL | Acrobot · Acrobot | performance relative to expert | Helped |
| Use AggreVaTeD instead of RL | CartPole · CartPole | performance relative to expert | Helped |
| Use an LSTM policy with AggreVaTeD in the partially observed setting | Acrobot (POMDP) · Acrobot (POMDP) | performance relative to expert | Helped |
| Use AggreVaTeD instead of the expert | Walker · Walker | performance relative to expert | Helped |
| Use AggreVaTeD instead of the expert | Hopper · Hopper | performance relative to expert | Hurt |
| Use regular gradient instead of natural gradient (ablation) | AggreVaTeD with an LSTM · Handwritten algebra dependency parsing | UAS | Mixed |
Closest papers by meaning
- Foundation Policies with Hilbert Representations (ICML 2024 · similarity 0.87)
- Causal Imitation for Markov Decision Processes: a Partial Identification Approach (NeurIPS 2024 · similarity 0.87)
- Diffusion-Reward Adversarial Imitation Learning (NeurIPS 2024 · similarity 0.87)
- Blending Imitation and Reinforcement Learning for Robust Policy Improvement (ICLR 2024 · similarity 0.87)
- Memory-Consistent Neural Networks for Imitation Learning (ICLR 2024 · similarity 0.86)
- Deep Policy Gradient Methods Without Batch Updates, Target Networks, or Replay Buffers (NeurIPS 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.