Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 20 of 84 for “"End-to-End Learning"”.
-
End-to-end learning of local point cloud feature descriptors
… technologies like augmented reality and autonomous vehicles have resulted in a growing need to identify and track objects in the environment. Object tracking and localization is frequently accomplished through the use of local feature descriptors, either in 2D or 3D. However, …
-
End-to-end learning, and audio-visual human-centric video understanding
<p>The field of machine learning has seen tremendous progress in the last decade, largely due to the advent of deep neural networks. When trained on large-scale labelled datasets, these machine learning algorithms can learn powerful semantic representations directly from the input data, end-to-end. …
-
Differentiable Minimum Snap Trajectory Planning for UAV Navigation with End-to-End Learning
This thesis presents a novel approach to robotic path planning, combining the efficiency of end-to-end learning with the smoothness of minimum snap trajectories. We introduce an enhanced version of Planner Net, a neural network architecture designed to generate optimal trajectories directly from …
-
What can we learn from touch? Towards end-to-end learning from robot touch
Physical interaction is fundamental to the human experience. Our embodiment has been optimised through aeons of evolution for the ability of general purpose physical interaction. Currently, there is a distinct gap between the physical capabilities of humans and those of robots. In robotics, rich …
-
End-to-end learning framework for circular RNA classification from other long non-coding RNAs using multi-modal deep learning.
… has become the focus of scientific studies due to its major role as a microRNA (miR) ac tivity modulator and its association with various diseases including cancer. Therefore, the detection of circular RNAs is a vital operation for continued comprehension of their biogenesis and purpose. …
-
Navigation and Localization for Railway Inspection Drone in GPS-denied Environments: An Investigative Study of Modular SLAM Baselines and End-to-End Learning-based Approaches
Autonomous navigation in GPS-denied environments remains a critical challenge for unmanned aerial vehicles performing infrastructure inspection. This thesis investigates the feasibility of learning-based navigation for railway-inspection drones by first constructing and analyzing a state-of-the-art …
-
DIFFERENTIABLE ROBOTICS: COMPOSITIONAL DEEP LEARNING WITH DIFFERENTIABLE ALGORITHM NETWORKS
Towards human-level robot intelligence a central question is the architecture of robot learning: how is the system represented, and how is it trained? This thesis introduces the Differentiable Algorithm Network (DAN), a compositional architecture for designing robot learning systems. The DAN is …
-
Deep generative models for phase retrieval in computational microscopy
Computational imaging is an important tool in science and medicine. It enables us to observe nano scale phenomenon which are beyond the range of any lens-based optical setup. We here tackle the problem of phase retrieval which occurs in a plethora of imaging tasks such as computational microscopy, …
-
Structure-aware machine learning over multi-relational databases
<p>We consider the problem of computing machine learning models over multi-relational databases. The mainstream approach involves a costly repeated loop that data scientists have to deal with on a daily basis: select features from data residing in relational databases using feature extraction …
-
Deep Dynamic Graphical Models for Molecular Kinetics
… a great demand for efficient analysis methods to distill essential information from simulation and enable quantitative characterization of molecular properties. Dynamic Graphical Model (DGM) is currently the most data-efficient method towards this goal. However, DGMs rely on extensive manual …
-
Reinforcement Learning with a Lost Person Model for Search and Rescue Path Planning
In this thesis, we train a reinforcement learning agent to plan paths for search and rescue applications using a model of lost person behavior trained on past search incidents. We propose an improved method for producing occupancy maps from the trajectories of an agent-based lost person model. We …
-
Deep Learning Methods for Nonlinearity Mitigation in Coherent Fiber-Optic Communication Links
… telecommunication services is rapidly growing. To meet this everincreasing connectivity demand telecommunication industry needs to maintain the exponential growth of capacity supply. One of the central efforts in this initiative is directed towards coherent fiber-optic communication systems, the …
-
Data-driven early prediction and modeling of lithium-ion battery capacity degradation
… a variety of Li-ion battery market participants to accurately model and predict the capacity degradation to improve the awareness of battery aging behaviors in making battery planning and operating decisions. However, there are two challenges in incorporating capacity degradation into the …
-
From Data to Decisions: Machine Learning for Enterprise Demand Forecasting
… This dissertation presents a machine learning (ML)-driven demand forecasting framework implemented at a Fortune-500 company HP Inc., focusing on three key areas: ML-based predictive modeling, MLOps and deployment scalability, and Human-in-the-loop forecasting integration. Additionally, …
-
Cardinality-Constrained Financial Optimization Problems
To construct an optimal portfolio, an investment process typically involves a risk model and an asset allocation. Whether it is which factors to include in the risk model or which assets to include in the portfolio, investors are constantly making a choice. If the selection is made a-priori, then …
-
SAMPLE EFFICIENT REPRESENTATION LEARNING FOR VISUAL RECOGNITION
… specifically focuses on the representation learning problem in practical scenarios and makes efforts to deal with imbalanced data distribution and unlabeled data. Our studies show that the commonly adopted joint (or end-to-end) learning scheme substantially harms the quality of the learned …
-
Dirbtinio intelekto kameros regėjimo sistema robotui automobiliui /
… domain, suchas computer vision, machine learning, autonomous driving and end-to-end learning. Theincrease of human need, the abundance of research, and the increased availability of com-puting power have also led to a better development in the robotics sector. As with any rapidlygrowing …
-
LiDAR Based Object Detection and Tracking in Stationary Applications
… are made for geometric feature-based methods and end-to-end learning methods for object detection from pointcloud data. Subsequently, theoretical pointcloud spacing models for multi-beam 360 deg LiDAR sensors are developed, with analysis on placement strategies and LiDAR configurations. The thesis …
-
Learning Rich Representations for 4D Scene Understanding
Autonomous systems operating in dynamic environments require a holistic understanding of 4D scenes — reasoning across 3D space and time — to safely interact with the world. Among various applications, autonomous vehicles have gained significant attention due to their projected societal impact. …
-
UNIVERSAL NAVIGATION: LEARNING A GENERAL NAVIGATION POLICY FOR HETEROGENEOUS ROBOTS
… is a challenging problem that requires a robot to find the goal using only visual inputs. Many researchers have demonstrated promising results using deep reinforcement learning (deep RL) on various robotic platforms, but typical end-to-end learning is known for its poor extrapolation capability …
Page 1 of 5