Co-authored-by: Remco Bloemen <[email protected]>

Readme.md

@@ -1,6 +1,25 @@
 # ZKP Neural Networks
 
-Prototype of evalutation neural networks in zero knowledge proofs using the plonky2 proof system.
+Prototype of evalutation of neural networks inside zero knowledge proofs using the plonky2 proof system.
+
+## ZKML
+
+To find out more about zero-knowledge machine learning, check out the [awesome-zkml](https://github.com/worldcoin/awesome-zkml) repository we have created. It aggregates scientific research papers, codebases, articles, and use cases in the field of ZKML.
+
+## Potential Worldcoin use cases
+
+[Worldcoin](https://worldcoin.org/) is a [Privacy-Preserving Proof-of-Personhood Protocol](https://worldcoin.org/the-worldcoin-protocol). ZKML could help us make our protocol more trustless, and make it more easily upgradeable and auditabie.  
+
+- Verifying that a user has created a valid and unique [WorldID](https://id.worldcoin.org) locally by running the IrisCode model on self-hosted biometric data and is calling [_addMember(uint256 groupId, uint256 identityCommitment)](https://github.com/semaphore-protocol/semaphore/blob/4e6be04729ed2d7e29461a3915877a66a2c9c4d2/contracts/base/SemaphoreGroups.sol#L43) function on the WorldID Semaphore identity group with a valid identityCommitment. -> Makes protocol more permissionless
+- Making the Orb trustless, provide proof that fraud filters on the hardware and firmware are applied
+- Enable IrisCode upgradeability
+
+## Technological stack
+
+- [Python](https://www.python.org/), [numpy](https://numpy.org/) - Flexible dynamic programming language and the fundamental package for scientific computing in Python -> used to create a vanilla implementation of a [CNN](https://en.wikipedia.org/wiki/Convolutional_neural_network)
+- [Rust](https://www.rust-lang.org/) - A performant, memory safe, systems-level programming language.
+- [plonky2](https://github.com/mir-protocol/plonky2): Powerful zero-knowledge proving system developed by the [Polygon Zero team](https://polygon.technology/solutions/polygon-zero/) -> to create zero knowledge circuits of the inference step of a neural network
+- [serde](https://serde.rs/) - Serialization/Deserialization library for Rust
 
 ## Build and run
 
@@ -19,7 +38,7 @@ python3 vanilla_cnn.py
 python3 generate_cnn_json.py
 cd ../
 # run Rust CNN implementation and compare results against your previous results
-cargo test nn::tests::serde_neural_net -- --show-output  
+cargo test serialize::tests::deserialize_nn_json -- --show-output  
 ```
 
 - This will run the vanilla CNN Python implementation and generate JSON files for the random matrices generated by numpy, these will be deserealized by serde in the Rust implementation and turned into  an ndarray `ArcArray<f32, IxDyn>`.
@@ -52,7 +71,7 @@ final output: [-0.11425511 -0.13403508 -0.41759714 -0.24778798  0.85626755]
 - Rust
 
 ```text
----- nn::tests::bench_neural_net stdout ----
+---- serialize::tests::deserialize_nn_json stdout ----
 layer                | output shape    | #parameters     | #ops           
 -----------------------------------------------------------------------------
 conv 32x5x5x3        | [116, 76, 32]      | 2400  | 7052800
@@ -77,7 +96,7 @@ cd ref_cnn
 python benchmark_cnn.py
 # generates matrices for the Rust implementation to use 
 python generate_cnn_json.py
-cargo bench bench_serde_neural_net
+cargo bench bench_neural_net
 ```
 
 ### Example output
@@ -93,7 +112,7 @@ The average time is 0.8297840171150046 seconds for 1000 runs
 - Rust: 0.151s
 
 ```text
-test nn::bench_serde_neural_net ... bench: 151,632,316 ns/iter (+/- 1,469,992)
+test nn::bench_neural_net ... bench: 151,632,316 ns/iter (+/- 1,469,992)
 ```
 
 In this benchmark the Rust implementation is **5.5x faster**!
@@ -109,5 +128,50 @@ cargo test
 In order to see output use `cargo test -- --output`, i.e.:
 
 ```text
-cargo test nn::tests::serde_neural_net -- --show-output 
+cargo test nn::tests::neural_net -- --show-output 
+```
+
+## Serialize/Deserialize CNN model
+
+### Python to JSON -> JSON to Rust
+
+Serializing the vanilla CNN model created with numpy into JSON and desearilizing the model into a `NeuralNetwork` Rust object
+
+```bash
+# change directory to cnn folder
+cd ref_cnn
+# generate json file for the model
+python generate_cnn_json.py
+
+cargo test serialize::tests::deserialize_model_json -- --show-output  
+```
+
+### Rust to JSON
+
+```bash
+# serializes a CNN model with random weights into src/json/nn.json
+cargo test serialize::tests::serialize_model_json -- --show-output  
+```
+
+### Full circle
+
+Create a `NeuralNetwork` object with random weights in Rust, serialize it into JSON and deserialize back into a `NeuralNetwork` Rust object
+
+```bash
+# serializes a CNN model with random weights into src/json/nn.json
+cargo test serialize::tests::serde_full_circle -- --show-output  
+```
+
+### Serialization/Deserialization benchmarks
+
+Benchmarks for serializing and deserializing the reference CNN (Rust/JSON) using [serde](https://serde.rs/).
+
+```bash
+# full serialization benchmark times (M1 Max Macbook Pro)
+# cargo bench - 579,057,637 ns/iter (+/- 20,202,535)
+cargo bench bench_serialize_neural_net
+
+# full deserialization benchmark times (M1 Max Macbook Pro)
+# cargo bench - 565,564,850 ns/iter (+/- 61,387,641)
+cargo bench bench_deserialize_neural_net         
 ```