Marco Altini - PhD, Data scientist, Entrepreneur
I'm leading data science activities at Bloom Technologies, a digital health startup focusing on helping expecting mothers have a healthy pregnancy. We are taking a data driven approach, combining data collected in clinical settings as well as consumer generated data to help shedding light on many poorly understood links between physiological changes naturally occurring during pregnancy, behavior and pregnancy outcomes.
I am also the creator of HRV4Training, a mobile platform using advanced signal processing and data analytics to optimize training. HRV4Training is the first and only app that can reliably measure heart rate variability (HRV) using the phone's camera, and uses HRV to optimize training programs and prevent overtraining. You can find out more on the HRV4Training blog.
In 2015 I obtained my PhD cum laude (top 5%) in applied machine learning at Eindhoven University of Technology. My PhD research at TU/e focused on applying machine learning techniques to develop new methods for personalized assessment of physical activity and cardiorespiratory fitness using wearable sensors data. During this period, I've published more than 25 peer reviewed papers for international conferences and journals, and three patents, which have been licensed to multiple customers in the high-end fitness industry. You can find most of my research here.
Previously, I obtained my M.Sc. degree cum laude in computer science engineering in 2010 from the University of Bologna. Between 2009 and 2014 I worked at imec as part of the Human++ program, developing hardware, firmware, software and algorithms for Body Area Networks applications.
Since 2012, I've been developing mobile apps using the phone's sensors to provide unique insights on the user's physiological status (such as heart rate and heart rate variability) without the need for external hardware. You can find out more in my blog.
A part from my work, I am interested in self-tracking, contemporary art, photography. I love running.
I am also the creator of HRV4Training, a mobile platform using advanced signal processing and data analytics to optimize training. HRV4Training is the first and only app that can reliably measure heart rate variability (HRV) using the phone's camera, and uses HRV to optimize training programs and prevent overtraining. You can find out more on the HRV4Training blog.
In 2015 I obtained my PhD cum laude (top 5%) in applied machine learning at Eindhoven University of Technology. My PhD research at TU/e focused on applying machine learning techniques to develop new methods for personalized assessment of physical activity and cardiorespiratory fitness using wearable sensors data. During this period, I've published more than 25 peer reviewed papers for international conferences and journals, and three patents, which have been licensed to multiple customers in the high-end fitness industry. You can find most of my research here.
Previously, I obtained my M.Sc. degree cum laude in computer science engineering in 2010 from the University of Bologna. Between 2009 and 2014 I worked at imec as part of the Human++ program, developing hardware, firmware, software and algorithms for Body Area Networks applications.
Since 2012, I've been developing mobile apps using the phone's sensors to provide unique insights on the user's physiological status (such as heart rate and heart rate variability) without the need for external hardware. You can find out more in my blog.
A part from my work, I am interested in self-tracking, contemporary art, photography. I love running.
updates
14/02/2016 - The last paper part of my PhD thesis was accepted for publication in the Journal of Applied Physiology. The paper's title is "Cardiorespiratory fitness estimation using wearable sensors: laboratory and free-living analysis of context-specific submaximal heart rates". In this work we use pattern recognition methods to contextualize HR in free living and predict fitness in unsupervised free-living settings, without requiring any laboratory protocol.
25/02/2016 - I was interviewed for the Quantified Body podcast, in which I cover a few aspects related to wearable technology, energy expenditure estimation and cardiorespiratory fitness estimation (mainly work that was part of my PhD). You can find the podcast here.
|
14/02/2016 - Published our research on estimating cardiorespiratory fitness in free living using wearable sensors in Artificial Intelligence in Medicine: M. Altini et al. "Cardiorespiratory fitness estimation in free-living using wearable sensors". See research & publication page.
11/02/2016 - We won the extreme tech challenge, organized by Richard Branson, with Bloom Technologies. Read more here.
|
15/01/2016 - Released HRV4Training Coach on the apple store. HRV4Training Coach provides access to all athletes' data & insights anytime, anywhere. You can find out more information here.
|
10/01/2016 - At Bloom Technologies we were selected as finalist of the Extreme Tech Challenge organized by Richard Branson. More info here.
|
15/12/2015 - I was nominated for the best reviewer award by the Journal of Biomedical and Health Informatics (JBHI).
15/12/2015 - Defended successfully my PhD at Eindhoven University of Technology (cum laude, top 5%). Slides available here.
|
10/12/2015 - Presented part of my work on heart rate variability and HRV4Training at the Running Data Science meetup, in Utrecht. Slides available here.
26/09/2015 - I was invited for the opening of TodaysArt where I gave a talk about my experience using HRV to better understand stress. I pretty much followed the story I wrote on medium. Before my talk, I also attended an expert meeting organized by Baltan Laboratories (Hack the Body), in the context of using sensors to understand how our body reacts to different stimuli and stressors. A great cross-disciplinary experience.
|
18/09/2015 - I gave two talks about heart rate variability at the Quantified Self Europe conference, in Amsterdam. The talks were mainly about tools & best practices for 60 seconds HRV measurements, as well as insights from the data in the context of monitoring training load, stress, and other factors influencing your physiology. You can find the slides here and here.
|
03/09/2015 - My PhD research is featured again in this month's issue of IEEE JBHI. In particular, the article was part of the special issue on the quantified self and sensors informatics, and covers my latest work on estimating personalized energy expenditure in the wild, by using multiple level of activity-derived context and contextualized heart rate. You can find the paper here.
|
18/08/2015 - I've been writing a series of blog posts on best practices for HRV measurements on the HRV4Training blog. In particular, I covered context & measurement time of the day, measurement duration, measurement type (lying, standing or orthostatic) and paced breathing.
|
01/08/2015 - Started working as freelance (on the side), mainly consulting around digital health technologies, physical activity monitoring, energy expenditure, heart rate variability and data science.
10/05/2015 - I released new versions of HRV4Training and Camera HRV. Both provide now a better user experience and a few additional features (integration with the Health app, population HRV comparisons, paced breathing, etc.). See itunes descriptions for more details.
|
28/03/2015 - In the final efforts of my PhD I am moving from laboratory based models to unsupervised free-living context recognition for personalized energy expenditure (EE) estimation. More specifically, I've developed a method to contextualize heart rate (HR) in free living using both low level atomic activities and high level contexts (discovered using topic models). Contextualized HR is then used to predict person-specific normalization parameters and personalize EE models without requiring individual calibration. The paper "Personalization of Energy Expenditure Estimation in Free Living Using Topic Models" was accepted for publication in IEEE JBHI, for a special issue on Sensor Informatics and Quantified Self.
22/03/2015 - My latest work on hierarchical Bayesian modeling for energy expenditure and cardiorespiratory fitness estimation was accepted for publication in the Journal of Biomedical Informatics. I also blogged about part of my research (excluding the free-living work currently under review), you can find my thoughts here.
|
17/03/2015 - I spent the last few weeks defining a new indicator of cardiovascular endurance and fitness, based on simple parameters that can be acquired with minimal effort using a mobile phone. The indicator is called Fitness Index, and the app I developed around this concept is StayFit. You can read more about it in my blog, here.
|
08/02/2015 - I wrote my first piece on Medium, it's about my experience as indie app developer. If you love to make things, code or are just thinking about getting started, you might find it interesting.
|
07/01/2015 - Three journal papers accepted for publication. The first one is the result of a collaboration with Pierluigi Casale on transfer learning. His work for BSN was extended for the IEEE Internet of Things journal. The second one is our first publication at Bloom, it covers wearable technology for healthier pregnancies, and will be published in the Proceedings of the IEEE. Finally, work related to my PhD on non-steady-state oxygen uptake was accepted for publication in IEEE JBHI.
17/12/2014 - Part of my PhD research is featured as research highlight on the next number of IEEE's Journal of Biomedical and Health Informatics. You can find it here. |
11/11/2014 - After a few months of hard work, we launched Bloom Technologies at the SDC in San Francisco, where we won whatstrending's startup pitch competition (video here). Bloom is a consumer health company focusing on helping expecting mothers have a healthy pregnancy. We are working on a unique platform, combining data from our own wearable, the first contraction patch, together with datastreams coming from other sensors and phones. In the long term, we aim at getting a better understanding of what parameters are key in predicting labor and what behavioral markers are related to pregnancy complications such as gestational diabetes, pre-eclampsia, etc. We are taking a consumer driven and data driven approach, which I personally believe can be very promising considering the poor understanding we currently have on human pregnancy. Find out more on medium.
01/10/2014 - Filed two patent applications around data analytics for context recognition, personalized messaging and detection of abnormalities based on contextualized physiological data.
01/07/2014 - Started working at Bloom Technologies as Head of Data Analytics, basically leading data science activities. Bloom is a digital health startup focusing on helping expecting mothers have a healthy pregnancy. We are taking a data driven approach, combining data collected in clinical settings as well as consumer generated data to help shading light on many poorly understood links between physiological changes naturally occurring during pregnancy, behavior and pregnancy outcomes.
|
16/06/2014 - Nominated for Best Paper Award at BSN 2014.
02/06/2014 - Journal paper accepted for publication in Physiological Measurement: "Personalizing Energy Expenditure Estimation Using Physiological Signals Normalization During Activities of Daily Living". This work extends personalization techniques I previously introduced for heart rate also to galvanic skin response and respiration rate.
02/06/2014 - Journal paper accepted for publication in Physiological Measurement: "Personalizing Energy Expenditure Estimation Using Physiological Signals Normalization During Activities of Daily Living". This work extends personalization techniques I previously introduced for heart rate also to galvanic skin response and respiration rate.
11/04/2014 - I was among the 5 finalists of the Armour39 Challenge. The winning team was led by Kevin Logan (MACSEA tld), who worked on recognizing different types of exercises using the A39 accelerometer. |
17/03/2014 - Journal paper accepted for publication in the IEEE Journal of Biomedical and Health Informatics: "Estimating Energy Expenditure using Body-Worn Accelerometers: a Comparison of Methods, Sensors Number and Positioning".
07/03/2014 - Journal paper accepted for publication in Methods of Information in Medicine: "Automatic Heart Rate Normalization for Accurate Energy Expenditure Estimation: An Analysis of Activities of Daily Living and Heart Rate Features".
24/12/2013 - Paper accepted for ACOMORE (PerCom): "Self-Calibration of Walking Speed Estimations Using Smartphone Sensors".
07/03/2014 - Journal paper accepted for publication in Methods of Information in Medicine: "Automatic Heart Rate Normalization for Accurate Energy Expenditure Estimation: An Analysis of Activities of Daily Living and Heart Rate Features".
24/12/2013 - Paper accepted for ACOMORE (PerCom): "Self-Calibration of Walking Speed Estimations Using Smartphone Sensors".
19/12/2013 - Released new versions of the HRV Logger and HRV4Training apps. Camera HRV also released on the app store. More info on the techniques used for the development can be found in my blog. |
23/11/2013 - I finished 11th/709 on my first competition on Kaggle (accelerometer biometric). |
02/11/2013 - Presented my work on personalized energy expenditure estimation and multi-sensor estimations at Wireless Health 2013, in Baltimore (US). For the third year in a row I arrived second at the awards.
21/08/2013 - Kicked off a project with artists Chris Salter, Maurizio Martinucci (aka TeZ) and Marije Baalman, in collaboration with Baltan Laboratories.
07/07/2013 - Presented my work on Body-weight normalized energy expenditure estimation at EMBC, in Osaka (JP). Slides here, paper here.
10/06/2013 - Two papers accepted for Wireless Health 2013. The full paper analyzes the impact of multiple physiological signals (heart rate, galvanic skin response, skin humidity and respiration) on activity type recognition and energy expenditure estimation, using single and activity-specific models. The demo paper concerns the implementation of the work presented at PH2013 on a mobile phone.
12/05/2013 - Talked about quantifying fitness at the Quantified Self Europe conference, in Amsterdam (NL). Slides available on slideshare.
06/05/2013 - Presented my work on Personalized Energy Expenditure Estimation at Pervasive Health, in Venice (IT). Paper and slides are available online.
13/04/2013 - During the partner days at Holst Centre, we demoed our new iPhone real-time activity recognition and energy expenditure estimation algorithms (bottom of this page) and showed the ActiBot, an activity lamp that changes color based on your activity level.
05/04/2013 - Paper "Body Weight-Normalized Energy Expenditure Estimation Using Combined Activity and Allometric Scaling Clustering" accepted for EMBC.
27/03/2013 - Paper "Personalizing Energy Expenditure Estimation Using a Cardiorespiratory Fitness Predicate" accepted for Pervasive Health.
19/03/2013 - Paper "Unsupervised Activity Clustering to Estimate Energy Expenditure with a Single Body Sensor" accepted for BSN.
01/03/2013 - the STRP project was successful. DJs wore the necklace during the live performance, pics and vids coming soon.
19/12/2012 - Patent application submitted: "Device and method for calculating cardiorespiratory fitness level and energy expenditure of a human being"
23/10/2012 - Nominated for Best Paper Award at Wireless Health 2012. You can find my presentation here, and my paper here.
27/07/2012 - Started a collaboration with Shanshan Chen (PhD student at University of Virginia) on signal processing for physical activity recognition and energy expenditure estimation.
21/08/2013 - Kicked off a project with artists Chris Salter, Maurizio Martinucci (aka TeZ) and Marije Baalman, in collaboration with Baltan Laboratories.
07/07/2013 - Presented my work on Body-weight normalized energy expenditure estimation at EMBC, in Osaka (JP). Slides here, paper here.
10/06/2013 - Two papers accepted for Wireless Health 2013. The full paper analyzes the impact of multiple physiological signals (heart rate, galvanic skin response, skin humidity and respiration) on activity type recognition and energy expenditure estimation, using single and activity-specific models. The demo paper concerns the implementation of the work presented at PH2013 on a mobile phone.
12/05/2013 - Talked about quantifying fitness at the Quantified Self Europe conference, in Amsterdam (NL). Slides available on slideshare.
06/05/2013 - Presented my work on Personalized Energy Expenditure Estimation at Pervasive Health, in Venice (IT). Paper and slides are available online.
13/04/2013 - During the partner days at Holst Centre, we demoed our new iPhone real-time activity recognition and energy expenditure estimation algorithms (bottom of this page) and showed the ActiBot, an activity lamp that changes color based on your activity level.
05/04/2013 - Paper "Body Weight-Normalized Energy Expenditure Estimation Using Combined Activity and Allometric Scaling Clustering" accepted for EMBC.
27/03/2013 - Paper "Personalizing Energy Expenditure Estimation Using a Cardiorespiratory Fitness Predicate" accepted for Pervasive Health.
19/03/2013 - Paper "Unsupervised Activity Clustering to Estimate Energy Expenditure with a Single Body Sensor" accepted for BSN.
01/03/2013 - the STRP project was successful. DJs wore the necklace during the live performance, pics and vids coming soon.
19/12/2012 - Patent application submitted: "Device and method for calculating cardiorespiratory fitness level and energy expenditure of a human being"
23/10/2012 - Nominated for Best Paper Award at Wireless Health 2012. You can find my presentation here, and my paper here.
27/07/2012 - Started a collaboration with Shanshan Chen (PhD student at University of Virginia) on signal processing for physical activity recognition and energy expenditure estimation.
02/07/2012 - Started my PhD in the ACTLab group at TU/e. My PhD research at TU/e focuses on applying machine learning learning techniques to develop methods and models to derive personalized assessment of physical activity and health markers, using data obtained with wearable sensors. My main supervisor is Dr. Oliver Amft.
|
06/06/2012 - Paper submitted to Wireless Health 2012 accepted.
02/05/2012 - I took part in an Art & Science workshop organized in Holst Centre. Danielle Roberts was the artist running the workshop, which was organized jointly by Baltan Laboratories and Holst Centre. You can read Danielle's thoughts here. Alessandra Saviotti wrote a nice overview of the collaborations between Holst Centre and Baltan, discussing other Art & Science workshops and this one in particular. You can read it on Digicult.
02/05/2012 - I took part in an Art & Science workshop organized in Holst Centre. Danielle Roberts was the artist running the workshop, which was organized jointly by Baltan Laboratories and Holst Centre. You can read Danielle's thoughts here. Alessandra Saviotti wrote a nice overview of the collaborations between Holst Centre and Baltan, discussing other Art & Science workshops and this one in particular. You can read it on Digicult.