MLSM at NeurIPS!

The 36th conference on Neural Information Processing Systems (NeurIPS) will be held in New Orleans from November 28th to December 9th.

Our group will be presenting an interesting paper entitled “Bayesian Active Learning with Fully Bayesian Gaussian Processes“. In this paper, we exploit the bias-variance trade-off inherent in a Gaussian process’ hyperparameters to design two new active learning acquisition functions: Bayesian Query-by-Committee and Query-by Mixture of Gaussian Processes. For six simulators, we improve the active learning performance by 12%! We will use these insights to efficiently analyze the effects of policies in traffic simulators, leading to greener aviation traffic 🌱✈.

Check below for more details!

New Ph.D. graduation in our group: congrats Daniele!

On September 6th, 2022, Daniele Gammelli successfully defended his Ph.D. thesis entitled “Learning and Control for Adaptive Transportation Systems.”

Congratulations, Daniele!

MLSM Group at trafikdage!

Five members of our MLSM group attended the 26th Trafikdage annual conference on traffic and transport in Aalborg, Denmark from 22-23 August 2022. Francisco Camara Pereira organized an executive session titled “AI in modelling and managing traffic” to review some of the latest developments in AI, modeling and traffic management, but also contrast with existing “classical” approaches. During the session, Filipe Rodrigues had a talk titled “Machine learning approaches for transport predictions problems: are we making progress in the right direction? ” where he discussed the importance of stationarity, strong baselines and benchmarks in transport prediction problems (Click here for the slides). In addition, our team presented/co-authored four interesting papers. Check out below for the full list of papers including the slides.

MLSM Group at trafikdage: 4 Papers and an executive…

The 26th Trafikdage annual conference on traffic and transport will be held in Aalborg, Denmark from 22-23 August 2022. Francisco Camara Pereira is organizing an executive session titled “AI in modelling and managing traffic” to review some of the latest developments in AI, modeling and traffic management, but also contrast with existing “classical” approaches. The session will take place on August 22nd, 11:30-13:00. During the session, Filipe Rodrigues will have a talk titled “Machine learning approaches for transport predictions problems: are we making progress in the right direction? ” where he will discuss the importance of stationarity, strong baselines and benchmarks in transport prediction problems.

Our group will also be presenting four interesting papers. Check out below for more details.

New member in our group: welcome Keren-Or!

Keren-Or recently joined our group as a Guest PhD. Student. She will be working on using system thinking methods to improve causality structure for transportation operations models, under the supervision of Professor Francisco Camara Pereira.

She is a PhD. Student at Technion (Haifa, Isreal), having a BSc degree in Industrial Engineering and an MSc in Information Systems and Statistics from the same institution.

We wish her the best of luck on her stay with us! 😊

MLSM Group at ICMC 2022: 6 papers

Four members of our MLSM group attended the 7th International Choice Modelling Conference (ICMC) in Reykjavik, Iceland from 23-25 May 2022. They presented/co-authored six interesting papers. Check out below for the full list of papers including the slides. Further slides will be added in due course.

New Ph.D. graduation in our group: congrats Valentino!

On April 26th, 2022, Valentino Servizi successfully defended his Ph.D. thesis entitled “Mining user transport behavior from smartphones.”

Congratulations, Valentino!

New members in our group: welcome Gabriel and Miguel!

Gabriel Valença is a visiting PhD candidate from Instituto Superior Técnico (IST) in Portugal. He is supervised by Professors Filipe Moura and Ana Sá from ITS. He will be working with Professor Carlos Lima Azevedo during his stay at DTU.

He mainly focuses on exploring how to allocate road space dynamically over the hours, days and seasons, especially when road space is underused. At DTU, he will evaluate the impacts of different scenarios of dynamic road space allocation.

Miguel Costa is also a guest PhD candidate from IST where he is supervised by Professors Manuel Marques and Filipe Moura. He will also be co-supervised by Professor Carlos Lima Azevedo from DTU.

His research focuses on the mapping and understanding of objective and subjective urban cycling safety. Miguel is combining authoritative and open-source data, together with machine learning approaches to analyze the risks cyclists face, and their perceptions of those risks, which ultimately aims to make cities safer for those who ride a bicycle.

We wish them a productive and fun time during their stay with us! 😊

New member in our group: welcome Lorena!

Lorena Torres Lahoz recently joined our group as a Ph.D. student within the eMOTIONAL Cities project, where she will focus on the inclusion of emotions in discrete choice model frameworks using machine learning and scenario discovery. She will be working under the supervision of Professors Carlos Lima Azevedo and Francisco Camara Pereira.

She has graduated from DTU and Universidad Politécnica de Valencia with a double MSc degree in Mathematical modeling and computation.

We wish her best of luck on her PhD journey! 😊

MLSM Group at ICMC 2022: 6 papers

The 7th International Choice Modelling Conference (ICMC) will be held in Reykjavik, Iceland from 23-25 May 2022. Our group will contribute with six interesting papers. Check out below for more details.