Restructured examples directory and slightly modified "LICENSE" to cl…

…arify copyrights of contributions.

LICENSE

@@ -3,6 +3,10 @@ From Hopfield layers:
 Copyright (c) 2020, Institute for Machine Learning, Johannes Kepler University Linz (Bernhard Schäfl)
 All rights reserved.
 
+All other contributions:
+Copyright (c) 2020 the respective contributors
+All rights reserved.
+
 From PyTorch:
 
 Copyright (c) 2016-     Facebook, Inc            (Adam Paszke)

README.md

@@ -89,9 +89,9 @@ Generally, the Hopfield layer is designed to be used to implement or to substitu
 
 The folder [examples](examples/) contains multiple demonstrations on how to use the <code>Hopfield</code> as well as the <code>HopfieldPooling</code> modules. To successfully run the contained [Jupyter notebooks](https://jupyter.org), additional third-party modules like [pandas](https://pandas.pydata.org) and [seaborn](https://seaborn.pydata.org) are required.
 
-- [Bit Pattern Set](examples/bit_pattern_demo.ipynb): The dataset of this demonstration falls into the category of <i>binary classification</i> tasks in the domain of <i>Multiple Instance Learning (MIL)</i> problems. Each bag comprises a collection of bit pattern instances, wheres each instance is a sequence of <b>0s</b> and <b>1s</b>. The positive class has specific bit patterns injected, which are absent in the negative one. This demonstration shows, that <code>Hopfield</code> and <code>HopfieldPooling</code> are capable of learning and filtering each bag with respect to the class-defining bit patterns.
+- [Bit Pattern Set](examples/bit_pattern/bit_pattern_demo.ipynb): The dataset of this demonstration falls into the category of <i>binary classification</i> tasks in the domain of <i>Multiple Instance Learning (MIL)</i> problems. Each bag comprises a collection of bit pattern instances, wheres each instance is a sequence of <b>0s</b> and <b>1s</b>. The positive class has specific bit patterns injected, which are absent in the negative one. This demonstration shows, that <code>Hopfield</code> and <code>HopfieldPooling</code> are capable of learning and filtering each bag with respect to the class-defining bit patterns.
 
-- [Latch Sequence Set](examples/latch_sequence_demo.ipynb): We study an easy example of learning long-term dependencies by using a simple <i>latch task</i>, see [Hochreiter and Mozer](https://link.springer.com/chapter/10.1007/3-540-44668-0_92). The essence of this task is that a sequence of inputs is presented, beginning with one of two symbols, <b>A</b> or <b>B</b>, and after a variable number of time steps, the model has to output a corresponding symbol. Thus, the task requires memorizing the original input over time. It has to be noted, that both class-defining symbols must only appear at the first position of a sequence. This task was specifically designed to demonstrate the capability of recurrent neural networks to capture long term dependencies. This demonstration shows, that <code>Hopfield</code> and <code>HopfieldPooling</code> adapt extremely fast to this specific task, concentrating only on the first entry of the sequence.
+- [Latch Sequence Set](examples/latch_sequence/latch_sequence_demo.ipynb): We study an easy example of learning long-term dependencies by using a simple <i>latch task</i>, see [Hochreiter and Mozer](https://link.springer.com/chapter/10.1007/3-540-44668-0_92). The essence of this task is that a sequence of inputs is presented, beginning with one of two symbols, <b>A</b> or <b>B</b>, and after a variable number of time steps, the model has to output a corresponding symbol. Thus, the task requires memorizing the original input over time. It has to be noted, that both class-defining symbols must only appear at the first position of a sequence. This task was specifically designed to demonstrate the capability of recurrent neural networks to capture long term dependencies. This demonstration shows, that <code>Hopfield</code> and <code>HopfieldPooling</code> adapt extremely fast to this specific task, concentrating only on the first entry of the sequence.
 
 ## Disclaimer
 
@@ -100,4 +100,8 @@ Some implementations of this repository are based on existing ones of the offici
 - The implementation of [HopfieldCore](modules/activation.py#L11) is based on the implementation of [MultiheadAttention](https://github.com/pytorch/pytorch/blob/b31f58de6fa8bbda5353b3c77d9be4914399724d/torch/nn/modules/activation.py#L771).
 - The implementation of [hopfield_core_forward](modules/functional.py#L8) is based on the implementation of [multi_head_attention_forward](https://github.com/pytorch/pytorch/blob/b31f58de6fa8bbda5353b3c77d9be4914399724d/torch/nn/functional.py#L3854).
 - The implementation of [HopfieldEncoderLayer](modules/transformer.py#L12) is based on the implementation of [TransformerEncoderLayer](https://github.com/pytorch/pytorch/blob/b31f58de6fa8bbda5353b3c77d9be4914399724d/torch/nn/modules/transformer.py#L241).
-- The implementation of [HopfieldDecoderLayer](modules/transformer.py#L88) is based on the implementation of [TransformerDecoderLayer](https://github.com/pytorch/pytorch/blob/b31f58de6fa8bbda5353b3c77d9be4914399724d/torch/nn/modules/transformer.py#L303).
+- The implementation of [HopfieldDecoderLayer](modules/transformer.py#L88) is based on the implementation of [TransformerDecoderLayer](https://github.com/pytorch/pytorch/blob/b31f58de6fa8bbda5353b3c77d9be4914399724d/torch/nn/modules/transformer.py#L303).
+
+## License
+
+This repository is BSD-style licensed (see [LICENSE](LICENSE)), except where noted otherwise.

examples/bit_pattern/auxiliary

@@ -0,0 +1 @@
+../../auxiliary