Added roadmap draft (#82)

* Added roadmap and community event

Co-authored-by: Bernhard Vogginger <[email protected]>

docs/source/_toc.yml

@@ -6,6 +6,8 @@ parts:
   - file: what
     title: What is it?
   - file: about
+  - file: roadmap
+  - file: events 
 - caption: User guide
   chapters:
   - file: usage

docs/source/about.md

@@ -4,6 +4,10 @@ The Neuromorphic Intermediate Representation (NIR) is a joint effort between res
 It is a product of many clever actors, thinking deeply about the interplay between computation and physics, and it would be incorrect to claim that NIR is owned by any single actor.
 Rather, NIR is a community effort that serves to simplify the development of neuromorphic algorithms and applications, and it is only possible because of the many contributors who put in years of work across many organizations and universities.
 
+```{admonition} NIR Community 
+Connect to our community via [Discord](https://discord.gg/JRMRGP9h3c) or one of our many [online events](events).
+```
+
 ## Design principles
 NIR is designed according to two founding principles:
 
@@ -12,14 +16,17 @@ NIR is designed according to two founding principles:
 2. Balance between productivity and correctness
     * We are entering an area of research where scientific giants have made many lasting marks. We are aware that we will not get everything right in the first go, but we are compelled to try. We are equally compelled to keep learning and improving, and the NIR spec will change as we grow.
 
+```{admonition} NIR Roadmap
+See more about the development and progress of NIR in our [roadmap](roadmap).
+```
+
 ## Connectivity 
 Each computational unit is a node in a static graph.
 Given 3 nodes $A$ which is a LIF node, $B$ which is a Linear node and $C$ which is another LIF node, we can define edges in the graph such as:
 
 ## Format
 The intermediate represenation can be stored as hdf5 file, which benefits from compression. 
 
-
 ## Authors
 Authors (in alphabetical order):
 * [Steven Abreu](https://github.com/stevenabreu7)

docs/source/events.md

@@ -0,0 +1,25 @@
+# NIR community events
+
+NIR is meant to be practical and valuable from day one.
+We are dedicated to interacting and paying back to the community both to (1) listen to the needs of the community and (2) to share our progress and learnings so that the entire community can benefit.
+Below, you will find a list of events that we are hosting or participating in where you can meet us and learn more about NIR.
+
+```{admonition} Would you like to invite us?
+Would you like to invite us to an event or workshop? Get in touch with us via [Discord](https://discord.gg/JRMRGP9h3c) or [GitHub](https://github.com/neuromorphs/nir).
+```
+
+## 2024
+
+* [Telluride Neuromorphic Cognition Engineering Workshop](https://sites.google.com/view/telluride-2024), workshop, **July** 
+  * A workshop on neuromorphic engineering and cognition. NIR will feature in the [Neuromorphic systems for space application](https://sites.google.com/view/telluride-2024/topic-areas-2024/spa24-neuromorphic-systems-for-space-applications) topic area.
+* [NICE workshop](https://flagship.kip.uni-heidelberg.de/jss/HBPm?mI=257&m=showAgenda&showAbstract=-1#9564), workshop, **April 26**
+  * A tutorial on NIR with on-site SpiNNaker2 hardware.
+* [Open-Neuromorphic workshop](https://open-neuromorphic.org/workshops/neuromorphic-intermediate-representation/), video, **February 5** 
+  * Video introducing NIR with practical examples on [Speck](https://www.synsense.ai/products/speck-2/) and [SpiNNaker2](https://spinnaker2.gitlab.io/) from several of the NIR core developers
+
+## 2023
+
+* [NIR preprint on arXiv](https://arxiv.org/abs/2311.14641), paper, **December** 
+  * The first publication of NIR, detailing the specification and several real-life experiments
+* [Telluride Neuromorphic Cognition Engineering Workshop](https://sites.google.com/view/telluride-2023/home), workshop, **July** 
+  * A workshop on neuromorphic engineering and cognition. This is where NIR was born.

docs/source/roadmap.md

@@ -0,0 +1,57 @@
+# NIR roadmap
+
+NIR is in its early stages but is developing fast.
+Here, we present an organizational roadmap for the immediate future of NIR governance, followed by a list of concrete projects that are currently developing NIR.
+
+```{admonition} Community involvement
+:class: tip
+We are dedicated to interacting and paying back to the community and we are continuously improving the accessibility of NIR by (1) releasing educational material, (2) updating our docs, and (3) engaging with the community in open events.
+
+For more information, see our [events](events) page and join the discussion on our [Discord](https://discord.gg/JRMRGP9h3c) server.
+```
+
+## Organizational roadmap
+
+With [the NIR preprint](https://arxiv.org/abs/2311.14641), we have demonstrated the feasibility of NIR.
+Not only as a scientific idea, but as a practical and valuable tool for the development of neuromorphic systems.
+However, NIR has been born in a research environment.
+To make it self-sustaining, we are preparing to launch a series of projects that will ensure the continued development of NIR, while continuing to increase its value to the community.
+
+The most important pillar of NIR is the joint value for both industrial and academic partners alike.
+Therefore, the first phase of our roadmap will be to reach out to potential partners and invite them to joint projects for which we can raise funds---either through commercial or public sources.
+The coming phases will revolve around the fundraising and organization of these projects.
+Eventually, we will kick off the projects and establish a steering group for NIR as the projects mature and further direction is needed.
+
+```
+ _______________       _______________      _______________      _______________ 
+|    Phase 1    | ->  |    Phase 2    | -> |    Phase 3    | -> |    Phase 4    | 
+|   May  2024   |     |  July  2024   |    |   Sep. 2024   |    |   Dec. 2024   | 
+|   Outreach    |     |  Fundraising  |    |   Kickoff     |    | Steering grp. | 
+|_______________|     |_______________|    |_______________|    |_______________|  
+```
+
+
+## Active projects
+
+The follwing projects are concrete intiatives that are working to improve NIR.
+Feel free to reach out to the project owners if you are interested in contributing or if you have any projects you would like to add.
+
+### Improved SpiNNaker 2 support
+**Duration**: August 2024 </br>
+**Owner**: [Bernhard Vogginger](https://github.com/bvogginger) </br>
+**Description**: We are working on improving the support for the SpiNNaker 2 platform. This includes speeding up SNN processing with the machine learning accelerator,  supporting more neuron models and streaming data. More information can be found at the [SpiNNaker 2 project milestone 5](https://gitlab.com/spinnaker2/py-spinnaker2/-/milestones/5).
+
+### Improved Lava & Loihi support
+**Duration**: October 2024 </br>
+**Owner**: [Steven Abreu](https://github.com/stevenabreu7/) </br>
+**Description**: We are working on improving the support for the Lava and Loihi platforms, focusing on neuron model support and applications.
+
+### Energy comparisons on different computational substrates
+**Duration**: Until July 2024 </br>
+**Owner**: [Sadasivan Shankar](https://profiles.stanford.edu/sadasivan-shankar) </br>
+**Description**: It would be helpful to understand exactly how much energy we can save by lifting a NIR graph to neuromorphic hardware. In this project, we are developing a mapping from NIR graph to energy efficiency, given a computational platform. That is, we can take a NIR graph and calculate a rough estimate of the energy consumption of the graph on a given platform. This will help us to understand the potential energy savings **before** it is deployed.
+
+### Additional neuron models and plasticity
+**Timeline**: Start October 2024 </br>
+**Owner**: [Jens Egholm Pedersen](https://github.com/jegp) </br>
+**Description**: We are working on adding additional neuron models and plasticity rules to NIR, including the Fitzhugh-Nagumo model. Plasticity will require additional primitives that can, themselves, modify primitive parameters.