CS MSc Thesis Presentation Day October 31!
Plats: E-huset E:2405 Glasburen
Kontakt: birger [dot] swahn [at] cs [dot] lth [dot] se
Spara händelsen till din kalender
Three MSc theses to be presented on October 31, 2019
October 31 is the day for coordinated master thesis presentations in Computer Science at Lund University, Faculty of Engineering. Three MSc theses will be presented.
Note to potential opponents: Register as an opponent to the presentation of your choice by sending an email to the examiner for that presentation (firstname [dot] lastname [at] cs [dot] lth [dot] se). Do not forget to specify the presentation you register for! Note that the number of opponents may be limited (often to two), so you might be forced to choose another presentation if you register too late. Registrations are individual, just as the oppositions are! More instructions are found on this page.
10:15-11.00 (E:2405 Glasburen)
Presenter: Linfeng Lu
Title: Time Series Prediction Using LSTM
Examiner: Flavius Gruian
Supervisors: Pierre Nugues (LTH), Lars Novak (Sinch), Jianhua Cao (Sinch)
Time series prediction and anomaly detection is important in many areas. LSTM is a gated RNN, which can create paths through time that have derivatives that neither explode nor vanish through connection weights that may change at each time step. In the thesis, we experiment three different models for time series prediction and anomaly detection on data from Sinch. One of the models does not contain LSTM encoder decoder framework, while the other two include the LSTM encoder decoder framework. We compare and analyze the experiment results and discuss which model performs better on the given dataset.
11:15-12:00 (E:2405 Glasburen)
Presenter: Daniel Myhrman
Title: Determining Room Occupancy with Machine Learning Techniques
Examiner: Jörn Janneck
Supervisors: Pierre Nugues (LTH), Fredrik Karlsson (Sony Mobile Communications AB)
Today, meeting and conference rooms may be too big or small compared to the actual amount of people who use them. In this thesis, the occupancy count has been estimated with the use of motion sensor, room and booking data. A neural network regression model has been developed, specifically a bidirectional long short-term memory architecture, and is used to predict the amount of occupants in rooms. Motion sensors placed in rooms can detect movement, but its data cannot directly be translated into occupancy count. To determine the occupancy, a neural network model has been developed and trained on manually collected ground truth data.
13:15-14:00 (E:2405 Glasburen)
Presenters: Ludvig Rappe & Samuel Johansson
Title: Transitioning from C to Rust in Media Streaming Development
Examiner: Björn Regnell
Supervisors: Elizabeth Bjarnason (LTH), Jonathan Karlsson (Axis Communications AB), Srimanta Panda (Axis Communications AB)
Rust uses static types, a strict compiler, and a unique ownership system, in order to make guarantees about memory safety without the use of a garbage collector and without sacrificing performance. This paper presents a case study of evaluating a programming language transition from C to Rust through the creation and usage of a Programming Language Transition Framework. We found the Programming Language Transition Framework to be a helpful tool for evaluating a programming language transition, but there is room for further improvements. We found Rust to provide a far superior developer experience, while still performing on par with C. However, Rust’s ownership system gives it a steep learning curve even for those with prior experience in programming. Given that the impact on performance is acceptable, and that build system integration can be done successfully, we deem it feasible to transition from C to Rust in this case.