Welcome to the official home of ShodhX, a student-founded research collective building open, verifiable tools and methods for machine intelligence and engineering diagnosis. We bridge theoretical neural networks with strict physics to provide domain-agnostic machine fault diagnostics.
- Abhimanyu Prasad : : Co-founder, Lead Developer (abhiprd2000@gmail.com)
- Tasfin Mahmud : Co-founder, Lead Developer (tasfinmahmud1@gmail.com)
- cansyd: A five-layer causal-neuro-symbolic framework for machine fault diagnosis. It independently verifies neural predictions against machine physics to guarantee high-integrity engineering decisions.
- purva: Core Python codebase for Dataset creation in the Bhojpuri NLP
- Physics-Informed Neural Networks : Infusing real-world mechanical constraints into neural architecture.
- Causal Inference: Transitioning from simple pattern recognition to strict fault diagnostics.
- Neuro-Symbolic AI: Unifying statistical learning with logical symbolic reasoning.
- NLP: NLP for Low resource languages and resource generation
We believe in open, verifiable science.
- Mailing List: Reach our team directly via shodhx@googlegroups.com.
- Collaborate: Check out our open issues or submit a pull request on our core repositories!
