Proposte di tesi

Per noi, la ricerca di giovani talenti appassionati delle ultime tecnologie è essenziale. Ci siamo impegnati nel consolidare la cooperazione e l’armonia all’interno del team.
Chi si unirà a noi incontrerà una piccola famiglia dove tutti si aiutano a vicenda, imparano e godono di tanti momenti informali con i colleghi.

 

Per noi è importante mantenere alta la motivazione dei nuovi arrivati attraverso il continuo sviluppo di nuove tecnologie e la loro applicazione su problemi reali. Lo sviluppo di una tesi di laurea è per noi un periodo importante per valutare le capacità e i talenti delle persone che recluteremo subito dopo la laurea. Per gli studenti del CdL Magistrale questa è una grande opportunità per lavorare su tecnologie all’avanguardia, applicate a problemi del mondo reale.

Partners

Addfor_politecnico_torino
Addfor_universita_torino
addfor_politecnico_milano
addfor_univerita_bologna
addfor_universita_trento

Tesi disponibili

Transformers-based models for portfolio management

[Available]

A recent publication “On Pre-Trained Image Features and Synthetic Images for Deep Learning“ uses real images as background for synthetically generated objects in order to create a dataset for Deep Learning algorithms.

[read more]

Trajectory forecasting with Transformer networks

[Available]

A recent publication “On Pre-Trained Image Features and Synthetic Images for Deep Learning“ uses real images as background for synthetically generated objects in order to create a dataset for Deep Learning algorithms.

[read more]

Transformers architectures for time series forecasting

[Available]

A recent publication “On Pre-Trained Image Features and Synthetic Images for Deep Learning“ uses real images as background for synthetically generated objects in order to create a dataset for Deep Learning algorithms.

[read more]

Deep learning models for anomaly detection

[Available]

A recent publication “On Pre-Trained Image Features and Synthetic Images for Deep Learning“ uses real images as background for synthetically generated objects in order to create a dataset for Deep Learning algorithms.

[read more]

Efficient Transformers architectures

[Available]

A recent publication “On Pre-Trained Image Features and Synthetic Images for Deep Learning“ uses real images as background for synthetically generated objects in order to create a dataset for Deep Learning algorithms.

[read more]

Human action recognition

[Available]

A recent publication “On Pre-Trained Image Features and Synthetic Images for Deep Learning“ uses real images as background for synthetically generated objects in order to create a dataset for Deep Learning algorithms.

[read more]

Explanation methods for anomaly detection models

[Available]

A recent publication “On Pre-Trained Image Features and Synthetic Images for Deep Learning“ uses real images as background for synthetically generated objects in order to create a dataset for Deep Learning algorithms.

[read more]

Image captioning with Transformers

[Available]

A recent publication “On Pre-Trained Image Features and Synthetic Images for Deep Learning“ uses real images as background for synthetically generated objects in order to create a dataset for Deep Learning algorithms.

[read more]

Self-supervised for Visual Representation Learning

[Available]

A recent publication “On Pre-Trained Image Features and Synthetic Images for Deep Learning“ uses real images as background for synthetically generated objects in order to create a dataset for Deep Learning algorithms.

[read more]

Efficient Extraction of motion vectors from h264/h265 streams

[Available]

A video encoder is a system that is able to transform a raw video (a sequence of uncompressed frames) to a more transferrable and storable format. Conversely a video decoder is a system that transforms that streams back to images. The pair (encoder, decoder) is usually called a video codec.

[read more]

Use of h264/h265 motion vectors as object motion estimation

[Available]

AVC and HEVC two of the most used codecs in video industry internally uses intra and extra frame motion vectors in order to achieve a better compression.

[read more]

Tesi non più disponibili

Synthetic object Dataset Augmentation

[Expired]

A recent publication “On Pre-Trained Image Features and Synthetic Images for Deep Learning“ uses real images as background for synthetically generated objects in order to create a dataset for Deep Learning algorithms.

[read more]

Capsule Networks as alternative to DCNN

[Expired]

Geoffrey Hinton released last month a document titled “Dynamic Routing Between Capsules“ and the entire Deep Learning community was shaken by this article.

[read more]

Algorithms for Multiple Object Tracking

[Expired]

The aim of this thesis is to define objective methods for measuring the performance of multiple object tracking algorithms in real datasets.

[read more]

Scene Classification in Video Streams

[Expired]

ResNet is a Deep Learning Network considered a reference for Image Classification. It is used in cases where a high classification accuracy is required. RL in conjunction with Deep Learning has obtained outstanding results in Atari video games, the Go board-game and a more complex environment like StarCraft II.

[read more]

Testing Algorithms for image & Video understanding

[Expired]

In the last few years, many algorithms with remarkable effectiveness for Object Detection have been published but still some comparative metrics haven’t been defined.

[read more]

Build a Chatbot for Interviews on IBM Watson

[Expired]

IBM Watson™ is a SaaS engine for the development of Cognitive Computing applications.

[read more]

Comparison of Reinforcement Learning Frameworks

[Expired]

Reinforcement Learning (RL) is a class of machine learning algorithms in which an agent interacts by trial-and-error in an environment.

[read more]

Safe Reinforcement Learning

[Expired]

Reinforcement Learning (RL) [1] is a class of machine learning algorithms in which an agent interacts by trial-and-error in an environment.

[read more]

Unsupervised/semi-supervised video classification

[Expired]

When there are hundreds or thousands of cameras producing video streams all day long it is very useful to have an algorithm that analyses such streams instead of a human.

[read more]

Deep Genomics: harnessing the power of Deep Neural Networks in the analysis of biomolecular data

[Expired]

The human genome project [1], an international scientific research project with the goal of determining the sequence of nucleotide base pairs that make up human DNA, lasted roughly 15 years and cost $5 billion (adjusted for inflation).

[read more]

Algorithm Optimization on Embedded & Server GPU

[Expired]

NVIDIA TensorRT5.1™ is an Inference Optimizer and runtime that allows low latency and high throughput for Deep-Learning applications.

[read more]

Model Based Reinforcement Learning

[Expired]

Reinforcement Learning (RL) [1] is a class of machine learning algorithms in which an agent interacts by trial-and-error in an environment.

[read more]

Inverse Reinforcement Learning for Autonomous Driving

[Expired]

Reinforcement Learning [1] (RL) is an emerging field of Artificial Intelligence (AI) that is giving extraordinary results in different applications.

[read more]

Explanation of Deep Neural Networks Predictions

[Expired]

Artificial Neural Networks are biologically-inspired programming paradigm which enables a computer to learn from observational data [1].

[read more]

Generative Adversarial Networks for Domain Adaptation between synthetic and real images

[Expired]

Learning from as little human supervision as possible is a major challenge in Machine Learning. In the context of computer vision, Deep Learning is a class of supervised learning algorithms that require a great amount of human-labeled images in order to be trained [1].

[read more]

Tesi concluse [ITA]

Tesi concluse [ENG]