File:oneshot3.jpg
Size of this preview: 800 × 264 pixels. Other resolution: 1,351 × 446 pixels.
Original file (1,351 × 446 pixels, file size: 69 KB, MIME type: image/jpeg)
From (Duan et al. 2017) Figure 3: Comparison of different conditioning strategies. The darkest bar shows the performance of the hard-coded policy, which unsurprisingly performs the best most of the time. For architectures that use temporal dropout, we use an ensemble of 10 different downsampled demonstrations and average the action
distributions. Then for all architectures we use the greedy action for evaluation.
File history
Click on a date/time to view the file as it appeared at that time.
| Date/Time | Thumbnail | Dimensions | User | Comment | |
|---|---|---|---|---|---|
| current | 22:00, 22 February 2018 | | 1,351 × 446 (69 KB) | Isucholu (talk | contribs) | From [https://papers.nips.cc/paper/6709-one-shot-imitation-learning.pdf (Duan et al. 2017)] Figure 3: Comparison of different conditioning strategies. The darkest bar shows the performance of the hard-coded policy, which unsurprisingly performs the bes... |
You cannot overwrite this file.
File usage
The following page uses this file:
