I had the pleasure of representing Webacy and presenting at EthCC for a second time, this year in Brussels, Belgium. The AI x crypto space has been super hot, so for this talk I focused on practical applications and techniques.
Watch the full video here:
https://ethcc.io/archive/Predicting-Risk-Leveraging-AI-to-Analyze-Smart-Contracts
Here are additional notes from the talk:
The number of deployed smart contracts is growing steadily. More smart contracts = more potential for hacks. We face two major challenges: the (increasing) complexity of smart contracts, and the limitations of manual analysis. Enter AI/ML.
GNNs help model complex structures. Great paper here: https://ijcai.org/proceedings/2020/0454.pdf…
softer intro here: https://distill.pub/2021/gnn-intro/
Transformers bring context-aware analysis. deep learning: https://nature.com/articles/s41598-023-47219-0…
vulnerability detection with NLP: https://ieeexplore.ieee.org/document/10420108
Federated Learning ensures data privacy while improving security. FL enabled by Smart Contracts: https://arxiv.org/pdf/2107.10243
At Webacy we’re using everything at our disposal to build the best safety network for the ecosystem. Join us, work with us, and help us build the safer web3 future together.