| @inproceedings{asseman2018learning, | |
| title = {Learning beyond simulated physics}, | |
| author = {Asseman, Alexis and Kornuta, Tomasz and Ozcan, Ahmet S.}, | |
| year = {2018}, | |
| booktitle = {Modeling and Decision-making in the Spatiotemporal Domain Workshop, | |
| 32nd Conference on Neural Information Processing Systems (NeurIPS 2018)}, | |
| address = {Montr\'eal, Canada}, | |
| url = {https://openreview.net/pdf?id=HylajWsRF7}, | |
| note = {Introduces the IBM Double Pendulum Chaotic Dataset | |
| (21 high-speed videos at 400 Hz, 480x480 px, with tracked marker positions | |
| and derived arm angles). Apparatus: pivot-to-first-datum arm 91 mm, | |
| first-to-second-datum arm 70 mm, three 19 mm fiducial markers (red=pivot, | |
| green=first datum, blue=second datum). Dataset released under | |
| CDLA-Sharing-1.0 via IBM Data Asset eXchange (DAX).} | |
| } | |