Using Adversarial Examples in Natural Language Processing
| Thesis title in Czech: | Využití adverzálních příkladů pro zpracování přirozeného jazyka |
|---|---|
| Thesis title in English: | Using Adversarial Examples in Natural Language Processing |
| Key words: | Neuronové sítě, Adverzální příklady, Zpracování přirozeného jazyka, Regularizace, Evaluace |
| English key words: | Neural networks, Adversarial examples, Natural language processing, Regularization, Evaluation |
| Academic year of topic announcement: | 2016/2017 |
| Thesis type: | diploma thesis |
| Thesis language: | angličtina |
| Department: | Institute of Formal and Applied Linguistics (32-UFAL) |
| Supervisor: | prof. Ing. Zdeněk Žabokrtský, Ph.D. |
| Author: | hidden - assigned and confirmed by the Study Dept. |
| Date of registration: | 10.03.2017 |
| Date of assignment: | 10.03.2017 |
| Confirmed by Study dept. on: | 23.03.2017 |
| Date and time of defence: | 07.09.2017 11:30 |
| Date of electronic submission: | 18.07.2017 |
| Date of submission of printed version: | 19.07.2017 |
| Date of proceeded defence: | 07.09.2017 |
| Opponents: | Mgr. Jindřich Libovický, Ph.D. |
| Guidelines |
| Deep neural networks have lately achieved state-of-the-art performance at many tasks. Nevertheless, even the leading models might be easily confused by artificially created examples. One of the newly developed training method relies on constructing such adversarial examples which are designed to cause a neural network to produce wrong outputs and which are consequently used for gradient update. The aim of this thesis is to explore applicability of this strategy in the field of Natural Language Processing by designing and evaluating experiments on a collection of various NLP datasets. |
| References |
| Haykin, Simon S., et al. Neural networks and learning machines. Vol. 3. Upper Saddle River, NJ, USA:: Pearson, 2009.
Rojas, Raúl. Neural networks: a systematic introduction. Springer Science & Business Media, 2013. LeCun, Yann, Yoshua Bengio, and Geoffrey Hinton. "Deep learning." Nature 521.7553 (2015): 436-444. Bengio, Yoshua, Ian J. Goodfellow, and Aaron Courville. "Deep learning." Nature 521 (2015): 436-444. |