Neural network library for natural language processing, implemented in F#. This library was used to obtain results in following papers:
- Tarasov D.S. (2015) Deep Recurrent Neural Networks for Multiple Language Aspect-Based Sentiment Analysis // Computational Linguistics and Intellectual Technologies: Proceedings of Annual International Conference “Dialogue”, Issue 14(21), V.2, pp. 65-74
- Tarasov D.S. (2015) Natural Language Generation, Paraphrasing and Summarization of User Reviews with Recurrent Neural Networks // Computational Linguistics and Intellectual Technologies: Proceedings of Annual International Conference “Dialogue”, Issue 14(21), V.1, pp. 571-579
Some features:
- Feedforward neural networks
- Uni and bi-directional multi-layer reccurrent networks
- Support for sequential and graph-based layer connectivity
- Sigmoid, Tanh, RelU and Softmax activations
- Dropout
- Projection layer for word embeddings
- Gaussian noise layer
- Runs under .NET Framework and mono runtime
- 1D convolutional layers and k-max pooling layers for text processing