Feature Request: Exploring WetML & Bio-Computing Systems Interfaces

[hpssjellis]

Feature Request: Exploring WetML & Bio-Computing Systems Interfaces (Cortical Labs CL1 / Maxwell Biosystems MaxOne style architectures)

Context & Motivation

As the scope of machine learning systems evolves beyond silicon, we are rapidly approaching an era where biological substrates will interface directly with computational pipelines. While platforms like Cortical Labs' Cl1 and MaxWell Biosystems' Maxone are primarily hardware testbeds today, integrating biological neural networks (BNNs) into machine learning training loops is becoming an inevitable frontier for systems engineering.

To help train the next generation of AI engineers to think about unconventional hardware substrates, it's worth planting a flag early on how ML systems will handle WetML (biocomputing interfaces).

Proposed Exploration / Discussion Topic

An architectural blueprint for Multi-Zone Bio-Network Simulation to Multi-Zone Trained Response Pipeline:

1. Input Zone: Encoding digital feature streams or spike trains into biological stimuli via high-density microelectrode arrays (MEAs).
2. Biological Processing Zone: Cultured neural networks acting as the dynamic, non-silicon intermediary handling wave propagation and plasticity.
3. Readout / Output Zone: Capturing multi-channel electrophysiological responses and feeding them back into traditional gradient-based or reinforcement learning layers on the edge/cloud side.

Why this matters for MLSysBook

• Broadening System Boundaries: Pushing systems thinking past pure silicon accelerators (CPUs/GPUs/NPUs) into wetware integration.
• Infrastructure Challenges: Dealing with extreme latency, biological drift, maintenance of living substrates, and closed-loop reinforcement on biological time scales.
• Future-Proofing Education: Giving students a conceptual framework for hybrid bio-silicon system engineering before it hits mainstream production.

Looking forward to thoughts on how we might introduce an exploratory section or issue brief on non-silicon/wetware ML infrastructure paradigms!

Resources (Needs updating as things change)

• MaxWell Biosystems — Manufacturer of the MaxOne (single-well) and MaxTwo (multi-well) HD-MEA platforms, widely cited for high-resolution, long-term neuronal network monitoring.
• Cortical Labs — Developer of the CL1 biological computer, which utilizes a closed-loop stimulation paradigm directly integrated with neurons on a silicon chip.
• 3Brain — Provider of the CorePlate™ 96W, a high-throughput HD-MEA system capable of simultaneous recording across all wells.
• Multi Channel Systems — Manufacturer of the Mesh MEA, specifically designed for 3D organoid electrophysiology, allowing recordings from within the structure.
• Axion BioSystems — Maker of the Maestro Pro and Maestro Edge platforms; standard equipment for automated, high-throughput closed-loop stimulation and neuronal recording assays.
• Blackrock Neurotech — Manufacturer of the Utah Array and high-density penetrating probes, primarily utilized for in-vivo neural interfacing.
• Neuronexus — Provides a wide range of silicon-based neural probes (planar and 3D) optimized for diverse in-vitro and in-vivo network recording applications.
• Cambridge NeuroTech — Specializes in ultra-dense silicon probes and specialized microelectrode arrays, highly regarded for signal-to-noise ratio in network-level recordings.
• Alpha MED Scientific — Originators of the MED64 system, offering established platforms for high-sensitivity recordings from stem-cell-derived models and brain slices.
• TissueLabs — Focused on organoid-compatible platforms and biofabrication, relevant for researchers integrating electrophysiology into custom organ-on-chip neural models.
• Harvard Bioscience — Distributor of the Mesh MEA technology (see Multi Channel Systems) for organoid-specific electrophysiology.
• BMSeed — In Vitro Research Tools with True In Vivo Relevance.

[Shashank-Tripathi-07]

Pretty interesting topic to think about @hpssjellis . Do you want this to be listed in the Book or in the other parts of the project ? I really like this frontier of technology personally as it aims to not only improve computing but also to improve human cognitive capacities in the longer-term.

[hpssjellis]

@Shashank-Tripathi-07 lets do a bit more research to see if it shoukd be in the book, I will see if I can get some people to comment who are actively in this field.

[Shashank-Tripathi-07]

Sure, before you proceed to literally build this section. I would like @profvjreddi to review this and share his advise as he's the best person to help you out on this adventure 😄