Quantum library with useful tools to help accelerate and optimise deep learning models

Research
1

Hello everyone, I’ve released an open-source project called QuantumAccel which is built around a symbolic logic engine that transforms traditional logic gates like AND, XOR, and Toffoli into optimised quantum-inspired operations, all within a constrained mathematical space.

Features:

  • Ultra-fast logic compression using sparse attention
  • Evolving symbolic gates that simulate Hadamard, CNOT, XNOR
  • Memory-efficient operation (as low as 4 KB for massive input)
  • Reversible logic operations for feature extraction, pattern recognition, and error detection

Use Cases:

  • Quantum simulation
  • Edge AI with kilobytes of RAM
  • Memory compression & logic acceleration
  • NLP/vision feature extraction without neural nets
    Link to the repo: GitHub - fikayoAy/quantum_accel
    This is part of a larger symbolic AI framework I’m building. Would love your feedback or contributions! Let me know if you’re interested in symbolic computation, quantum logic, or memory-efficient learning.

Demo benchmarks and documentation are available in the repo. Apache Licensed.

2 Likes
2

Hey I forked your repo, your research is quite interesting, it’s exciting to see more and more people advance symbolic research daily.

I’m a psychology researcher and developer. I’m part of a team of researchers and engineers exploring Evolutionary AI, with a focus on bridging recursive reasoning, symbolics, and frontier ML attention architectures.

Here’s one of our projects on symbolics, as well the preprints we submitted to NeurIPS. Let me know if youd to like to discuss any topics further:

Symbolic Residue Diagnostic Suite: Tracks and diagnoses transformer model failure modes: silent inconsistencies or “residues” in reasoning paths. The structural data vectors behind why advanced reasoning fails.

https://github.com/recursivelabsai/Symbolic-Residue

1 Like
3

Hi, I’m really fascinated by your project, as I have never thought in that direction before… All I did was to compute the loss function and accept the loss as it was but not actually find out why and what exactly my model encountered and gave the result… And thank you for finding my research interesting. If you want to chat me up for collaboration or any interest as regards to this, please let me know. My Discord ID: 921787248591130694

2 Likes