From fd5bdcacdf3d82f1665074c6cb51dae2f3acb645 Mon Sep 17 00:00:00 2001 From: happyappledog Date: Mon, 6 Nov 2023 15:24:00 -0500 Subject: [PATCH 1/2] Added a medical example in AI for Good chapter --- ai_for_good.qmd | 7 +++++++ references.bib | 11 +++++++++++ 2 files changed, 18 insertions(+) diff --git a/ai_for_good.qmd b/ai_for_good.qmd index 33b946ac..669c1735 100644 --- a/ai_for_good.qmd +++ b/ai_for_good.qmd @@ -100,6 +100,13 @@ A collaborative research team from the University of Khartoum and the ICTP is ex This portable, self-contained system shows great promise for entomology. The researchers suggest it could revolutionize insect monitoring and vector control strategies in remote areas. By providing cheaper, easier mosquito analytics, TinyML could significantly bolster malaria eradication efforts. Its versatility and minimal power needs make it ideal for field use in isolated, off-grid regions with scarce resources but high disease burden. +### TinyML Design Contest (TDC) - TinyML contest in healthcare +The first TinyML contest in healthcare, TDC’22 [@jia2023life] was held in 2022 to motivate participating teams to design AI/ML algorithms for detecting life-threatening ventricular arrhythmias (VAs) and deploy them on Implantable Cardioverter Defibrillators (ICDs). VAs are the main cause of sudden cardiac death (SCD). People at high risk of SCD rely on the ICD to deliver proper and in-time defibrillation treatment (i.e., shocking the heart back into normal rhythm) when experiencing life-threatening VAs. On-device algorithm for early and in-time life-threatening VA detection will increase the chances of survival. + +The proposed AI/ML algorithm needs to be deployed and executed on an extremely low-power and resource-constrained microcontroller (MCU) (a $10 development board with an ARM Cortex-M4 core at 80 MHz, 256 kB of flash memory and 64 kB of SRAM). The submitted designs were evaluated by metrics measured on the MCU: (1) detection performance; (2) inference latency; and (3) memory occupation by the program of AI/ML algorithms. The champion, GaTech EIC Lab, obtained 0.972 in $F_β$ (F1 score with a higher weight to recall), 1.747 ms in latency and 26.39 kB in memory footprint with a deep neural network. + +An ICD with on-device VA detection algorithm was inplanted in a clinical trial (see the author's [presentation](https://youtu.be/vx2gWzAr85A?t=2359)). + ## Science In many scientific fields, researchers are limited by the quality and resolution of data they can collect. They often must infer the true parameters of interest indirectly, using approximate correlations and models built on sparse data points. This constrains the accuracy of scientific understanding and predictions. diff --git a/references.bib b/references.bib index 4501f577..9c36b8db 100644 --- a/references.bib +++ b/references.bib @@ -690,6 +690,17 @@ @inproceedings{altayeb2022classifying year={2022} } +@article{jia2023life, + title={Life-threatening ventricular arrhythmia detection challenge in implantable cardioverter--defibrillators}, + author={Jia, Zhenge and Li, Dawei and Xu, Xiaowei and Li, Na and Hong, Feng and Ping, Lichuan and Shi, Yiyu}, + journal={Nature Machine Intelligence}, + volume={5}, + number={5}, + pages={554--555}, + year={2023}, + publisher={Nature Publishing Group UK London} +} + @article{yamashita2023coffee, title={Coffee disease classification at the edge using deep learning}, author={Yamashita, Jo{\~a}o Vitor Yukio Bordin and Leite, Jo{\~a}o Paulo RR}, From 40c0c2eaaf7cec822e3b7f612e6f789253c87c16 Mon Sep 17 00:00:00 2001 From: Vijay Janapa Reddi Date: Tue, 7 Nov 2023 07:48:51 -0500 Subject: [PATCH 2/2] Reworked the flow a bit --- ai_for_good.qmd | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/ai_for_good.qmd b/ai_for_good.qmd index 669c1735..dec411e1 100644 --- a/ai_for_good.qmd +++ b/ai_for_good.qmd @@ -100,12 +100,13 @@ A collaborative research team from the University of Khartoum and the ICTP is ex This portable, self-contained system shows great promise for entomology. The researchers suggest it could revolutionize insect monitoring and vector control strategies in remote areas. By providing cheaper, easier mosquito analytics, TinyML could significantly bolster malaria eradication efforts. Its versatility and minimal power needs make it ideal for field use in isolated, off-grid regions with scarce resources but high disease burden. -### TinyML Design Contest (TDC) - TinyML contest in healthcare -The first TinyML contest in healthcare, TDC’22 [@jia2023life] was held in 2022 to motivate participating teams to design AI/ML algorithms for detecting life-threatening ventricular arrhythmias (VAs) and deploy them on Implantable Cardioverter Defibrillators (ICDs). VAs are the main cause of sudden cardiac death (SCD). People at high risk of SCD rely on the ICD to deliver proper and in-time defibrillation treatment (i.e., shocking the heart back into normal rhythm) when experiencing life-threatening VAs. On-device algorithm for early and in-time life-threatening VA detection will increase the chances of survival. +### TinyML Design Contest in Healthcare -The proposed AI/ML algorithm needs to be deployed and executed on an extremely low-power and resource-constrained microcontroller (MCU) (a $10 development board with an ARM Cortex-M4 core at 80 MHz, 256 kB of flash memory and 64 kB of SRAM). The submitted designs were evaluated by metrics measured on the MCU: (1) detection performance; (2) inference latency; and (3) memory occupation by the program of AI/ML algorithms. The champion, GaTech EIC Lab, obtained 0.972 in $F_β$ (F1 score with a higher weight to recall), 1.747 ms in latency and 26.39 kB in memory footprint with a deep neural network. +The first TinyML contest in healthcare, TDC’22 [@jia2023life], was held in 2022 to motivate participating teams to design AI/ML algorithms for detecting life-threatening ventricular arrhythmias (VAs) and deploy them on Implantable Cardioverter Defibrillators (ICDs). VAs are the main cause of sudden cardiac death (SCD). People at high risk of SCD rely on the ICD to deliver proper and timely defibrillation treatment (i.e., shocking the heart back into normal rhythm) when experiencing life-threatening VAs. -An ICD with on-device VA detection algorithm was inplanted in a clinical trial (see the author's [presentation](https://youtu.be/vx2gWzAr85A?t=2359)). +An on-device algorithm for early and timely life-threatening VA detection will increase the chances of survival. The proposed AI/ML algorithm needed to be deployed and executed on an extremely low-power and resource-constrained microcontroller (MCU) (a $10 development board with an ARM Cortex-M4 core at 80 MHz, 256 kB of flash memory and 64 kB of SRAM). The submitted designs were evaluated by metrics measured on the MCU for (1) detection performance; (2) inference latency; and (3) memory occupation by the program of AI/ML algorithms. + +The champion, GaTech EIC Lab, obtained 0.972 in $F_\beta$ (F1 score with a higher weight to recall), 1.747 ms in latency and 26.39 kB in memory footprint with a deep neural network. An ICD with an on-device VA detection algorithm was [implanted in a clinical trial](https://youtu.be/vx2gWzAr85A?t=2359). ## Science