Daniel Saromo-Mori
Background
Hello there! My name is Daniel and I am a M.Sc. student of the Automation and Control Engineering program at the Politecnico di Milano (POLIMI) (13th best engineering university in the world - QS World University Rankings).
On the other hand, I am a research engineer and teacher at the Artificial Intelligence Research Group (IA-PUCP). Also, I am a former research assistant at the Biomechanics and Applied Robotics Research Lab (GIRAB-PUCP). Both organizations are located at the Pontifical Catholic University of Peru (PUCP), where I received my B.Sc. and my Professional Engineer Degree in Mechatronic Engineering. Additionally, I was a guest researcher at the Bremen branch of the German Institute of Artificial Intelligence (DFKI).
My research advisors are Dr. Elizabeth Villota, Dr. Edwin Villanueva, and Dr. Matias Valdenegro-Toro. I am very thankful for their continuous support and guidance.
Experience in AI Research
I am the inventor of the Auto-Rotating Neural Networks, which are a generalization of the Auto-Rotating Perceptrons (ARP) I created in 2019. When using deep neural networks, do we ever change the neuron itself? The ARP neural unit changes the internal core of classic perceptrons of dense layers aiming to tackle the vanishing gradient problem (VGP).
Furthermore, as part of my job at the GIRAB-PUCP laboratory, I have used the ARP for calibrating a wearable sensor, greatly improving the performance of dense neural networks with classic perceptrons.
On the other hand, the ARP have been presented in six countries (Austria, Canada, Guatemala, Peru, Poland, and Singapore). Also, I have released the open-source ARP Library for Keras, which you can check here.
The ARP operation can also be applied to recurrent and convolutional layers!
Currently, I am working on the journal version of the ARP paper and pushing its limits. There, I extended the ARP concept and created the Auto-Rotating Neural Networks. It was hard work, but I have successfully implemented the Auto-Rotating version of the following neural network layer types: dense, recurrent, Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and convolutional layers. When compared with their classical versions, the experimental results obtained with the Auto-Rotating Neural Networks are promising. With the help of Dr. Valdenegro-Toro, we managed to elaborate and submit the ARNN paper to the Transactions of Machine Learning Research (TMLR) journal, where is currently under double-blind review. Our research paper is called Auto-Rotating Neural Networks: An Alternative Approach for Preventing Vanishing Gradients.
My work as a teacher
I enjoy teaching ☺. Besides teaching at PUCP, I have worked as a private contractor for the Peruvian government, giving the following AI-related courses:
- Introduction to Artificial Intelligence and Machine Learning for National Meteorological and Hydrological Services: Special course given to SENAMHI, the Peruvian government entity in charge of the Meteorology and Hydrology, which wanted to apply AI and ML to their internal processes.
- Python and R for Data Science applied to healthcare: Special course given to the Peruvian government agency EsSalud, which wanted to train their staff in the topics of ML, to improve their internal working processes.
Also, I have been a private tutor in subjects related to electronics and programming, for undergraduate students from PUCP, UNMSM, and UTEC.
On the other hand, I have been invited by the AEDITIP group to teach courses related to AI.
To infinity and beyond!
Currently, I am studying a M.Sc. in Milan. It’s a new experience and I am learning a lot every day, and growing up as a person. Sometimes things are difficult, and you could feel homesick.
If you arrived until here with your reading, here’s a little piece of advice that nobody asked for, but it helps me a lot: if you feel overwhelmed just do your BEST!: Take a step Back, Envision yourself now and how you would like to do to improve your situation, then Say it out loud (or in your mind), and then Take action to get a step closer to the life you want to live.
Contact: danielsaromo [at] gmail [dot] com
Research Interests
AI-powered robotics. Robot learning. Human-robot cooperation. Data-driven control. Auto-Rotating Neural Networks and Auto-Rotating Perceptrons. Deep Learning.
News
| Date | Description | Location |
|---|---|---|
| June 19-21, 2022. | Attended CVPR 2022. Presenting the paper “A Deeper Look into Aleatoric and Epistemic Uncertainty Disentanglement” in the LXCV workshop as an oral exposition and poster presentation. | Louisiana, USA. |
| April 23, 2022. | LXAI Travel Grant (for attending CVPR 2022 to be an oral and poster presenter). | Louisiana, USA. |
| March 10-13, 2022. | Attended the International Women’s Day Hackathon 2022, organized by the group Teens in AI. I was invited to be a mentor of one of the Peruvian teams, in the areas of AI and ML. My group was awarded first place in the hackathon. | Lima, Peru*. |
| November 8-19, 2021. | Attended the Asian Conference in Machine Learning (ACML) 2021. Presenting the poster “Auto-Rotating Neural Networks” in the Online Asian Machine Learning School (OAMLS). | Singapore, Singapore*. |
| October 01, 2021. | I was invited to be a speaker at the VII Research Meeting of the PUCP Engineering Department, where I presented my research on the application of my ARP to the calibration of a wearable sensor. | Lima, Peru*. |
| August 25-25, 2021. | Attended the International Meeting on Artificial Intelligence and its Applications (RIIAA) Summer School 2021. | - |
| July 4, 2021. | I was invited by the PUCP Student Group of Environment and Water Resources (GEMRA PUCP) to give an introductory workshop to MATLAB. GitHub repository available here. | Lima, Peru*. |
| June 28-30, 2021. | I was invited by the IEEE Engineering in Medicine & Biology Society PUCP (EMBS PUCP) to give the tutorial titled: “Introduction to Python and MicroPython”. GitHub repository available here. | Lima, Peru*. |
| March 22, 2021. | I gave a tutorial about the Auto-Rotating Perceptrons for the group Papers We Love Guatemala. In this talk, I taught how to use the ARP by presenting the Keras ARP Library and a example notebook. | Guatemala, Guatemala*. |
| December 16, 2020. | I was invited to give an oral presentation at the final class of the Machine Learning course at PUCP. I talked about my research in deep learning and showed my progress with the Auto-Rotating Perceptrons. | Lima, Peru*. |
| December 9, 2020. | I took the TOEFL iBT examination and obtained a score of 97/120. | - |
| July 12-18, 2020. | Attended ICML 2020. Presenting the paper “Smart Sensor Calibration with Auto-Rotating Perceptrons” in the LXAI workshop as an oral exposition and poster presentation. | Vienna, Austria*. |
| July 10, 2020. | I was invited by the PUCP Association of Mechatronics Engineering Students (ASIME PUCP) to give the talk titled: ”Getting to know your career: Mechatronics Engineering”. Timestamp: 2:27:42. | Lima, Peru*. |
| May 31, 2020. | Attended the International Conference on Robotics and Automation (ICRA) 2020. | Paris, France*. |
| February 29, 2020. | I was invited by the PUCP Federated Center of Science and Engineering Faculty (CEFACI PUCP) to give the talk titled: ”Myths and Truths: Science and Engineering - Mechatronics Engineering Edition”. | Lima, Peru. |
| December 8-14, 2019. | Attended NeurIPS 2019. Presenting the paper “Auto-Rotating Perceptrons” in the LXAI workshop as an oral exposition and poster presentation. | Vancouver, Canada. |
| November 9, 2019. | Innovation Recognition Award at ASME Old Guard 63rd Anual Oral Competition Finals at ASME IMECE 2019, where I presented my bachelor thesis project. | Utah, USA. |
| October 12, 2019. | LXAI Travel Grant (for attending NeurIPS 2019 to be an oral and poster presenter). | Vancouver, Canada. |
| August 14, 2019. | ASME Travel Award to represent PUCP and South America in the ASME IMECE 2019 Finals Competition. | Utah, USA. |
| August 9, 2019. | Won the first place and, also, the Technical Award at ASME Old Guard Oral Presentation Competition (ASME E-FEST South America 2019). | Lima, Peru. |
| July 16, 2019. | Best undergraduate thesis and poster presentation at the PUCP Mechatronics Engineering End of Career Workshop 2019. | Lima, Peru. |
| November 9, 2018. | Accepted into the Udacity’s PyTorch Scholarship Challenge From Facebook, where I was taught how to build, train, and deploy state-of-the-art deep learning models with PyTorch. | - |
| July 17, 2018. | Oral presenter of the project: Automatic system for pre-fried potatoes production. My team and I managed to design and implement a device that received washed potatoes for later processing. The machine then peeled, cut, and pre-fried the potatoes. It was awarded as the best project of the PUCP Mechatronics Project Fair 2018. | Lima, Peru. |
| October 6, 2017. | 11th place (national level) at the FESTO’s X Academic Mechatronics Olympics. | Lima, Peru. |
| August 02-05, 2016. | Attending INTERCON 2016. | Piura, Peru. |
* Held virtually due to the global COVID-19 situation.
