CORE Seminar - October 26, 2026 - Dr. Amir Hooshiar

Value of the Information & Strength of the Argument

The talk provides substantial value by presenting concrete, actionable solutions to real clinical problems: the Holopod addresses data capture, the prehabilitation model predicts outcomes, and the AI systems offer real-time guidance and feedback. The argumentation is strong: the speaker explains the rationale behind each approach, acknowledges limitations (e.g., data annotation bottleneck, regulatory constraints), and emphasizes the need for explainable AI. The presentation is persuasive, moving from a high-level vision to specific implementations with measured accuracies (e.g., >80% for outcome prediction, 78% for CVMPO detection), and it coherently ties together the three pillars. The argument that AI should augment rather than replace surgeons is well-supported by the emphasis on human-in-the-loop systems and expert-level care scaling.

Scientific Rigor, Source Quality, Title Accuracy

The speaker references authoritative sources such as the Royal College of Surgeons of England report and mentions collaborations with Nvidia, Google, and Microsoft. However, no direct citations or URLs are provided in the presentation, and the talk lacks formal peer-reviewed references. The title ‘CORE Seminar’ is non-descriptive, but the description clearly outlines the topic, so the title-content fit is adequate. The presentation is scientifically rigorous in its methods: the speaker discusses data acquisition challenges, federated learning for privacy, and explainability for regulatory approval. The use of specific metrics (accuracy percentages) and the demonstration of real-time system behavior add credibility. Nonetheless, as a seminar, it serves as an expert opinion and research overview rather than a systematic literature review or meta-analysis.

The title is generic (seminar listing) but the description clearly states the topic on AI in surgery. The content matches the description.

8/10

The speaker is a director of a surgical robotics center with FDA-approved devices, presents specific data and collaborations. The talk is rigorous and transparent about limitations, but lacks peer-reviewed citations in the presentation itself.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 02:00Introduction on the future of surgery as a moving target, citing Royal College of Surgeons' attributes.
• 07:00Clarification that surgical robotics aim to augment surgeons, not replace them; presentation of lab's logo and structure.
• 18:00Discussion of the data problem: 137 TB/day generated, 97% not captured; introduction of Holopod device.
• 25:00Prehabilitation platform: multimodal data encoders, outcome prediction >80% accuracy.
• 34:00Mock surgical robotic system and real-time AI for instrument recognition and phase detection.
• 41:00Explainability heat maps and attention visualization for surgical expertise assessment.
• 45:00Surgical phase recognition combined with expertise prediction for trainee feedback.
• 55:00CVMPO detection in hernia repair, demonstrating real-time guidance for junior surgeons.

• Future of Surgery report - Royal College of Surgeons of England — Cited at the beginning as a reference for the expected attributes of future surgery, including robotics and augmented reality.
• NSQIP (National Surgical Quality Improvement Program) — Mentioned as inspiration for the types of surgical outcome measures predicted by the prehabilitation models.