Axel Krieger (Inner Logic): Robots That Learn Surgery From Surgeons

Surgical robotics' next decade of value creation shifts from hardware to imitation-learning-based autonomy software, per Inner Logic's founder.

• Imitation learning replaces hand-coded control: After hitting a ~70% ceiling with model-based stitching, Axel Krieger's lab pivoted to imitation learning using Chelsea Finn's ACT (action chunking transformer) framework, letting robots learn directly from expert surgeons rather than rule-based automation.
• Eight-for-eight autonomous surgery phase: Trained on ~17 hours of data (30 ex-vivo pig/butcher-sourced gallbladders, ~15 reps each), the robot performed the clip-and-cut phase of gallbladder surgery eight times in a row with zero human intervention; results published in Science Robotics.

Themes, sentiment and reasons are AI-written paraphrases, checked so they never quote an episode. They may contain errors. Not investment advice. Episodes remain the property of their creators.