Theses and Dissertations
Filtering by
- All Subjects: Active learning
- All Subjects: Software Engineering
- Creators: Li, Baoxin
Speedsolving, the art of solving twisty puzzles like the Rubik's Cube as fast as possible, has recently benefitted from the arrival of smartcubes which have special hardware for tracking the cube's face turns and transmitting them via Bluetooth. However, due to their embedded electronics, existing smartcubes cannot be used in competition, reducing their utility in personal speedcubing practice. This thesis proposes a sound-based design for tracking the face turns of a standard, non-smart speedcube consisting of an audio processing receiver in software and a small physical speaker configured as a transmitter. Special attention has been given to ensuring that installing the transmitter requires only a reversible centercap replacement on the original cube. This allows the cube to benefit from smartcube features during practice, while still maintaining compliance with competition regulations. Within a controlled test environment, the software receiver perfectly detected a variety of transmitted move sequences. Furthermore, all components required for the physical transmitter were demonstrated to fit within the centercap of a Gans 356 speedcube.
depicted commendable performance in a variety of applications. A fundamental challenge
in training deep networks is the requirement of large amounts of labeled training
data. While gathering large quantities of unlabeled data is cheap and easy, annotating
the data is an expensive process in terms of time, labor and human expertise.
Thus, developing algorithms that minimize the human effort in training deep models
is of immense practical importance. Active learning algorithms automatically identify
salient and exemplar samples from large amounts of unlabeled data and can augment
maximal information to supervised learning models, thereby reducing the human annotation
effort in training machine learning models. The goal of this dissertation is to
fuse ideas from deep learning and active learning and design novel deep active learning
algorithms. The proposed learning methodologies explore diverse label spaces to
solve different computer vision applications. Three major contributions have emerged
from this work; (i) a deep active framework for multi-class image classication, (ii)
a deep active model with and without label correlation for multi-label image classi-
cation and (iii) a deep active paradigm for regression. Extensive empirical studies
on a variety of multi-class, multi-label and regression vision datasets corroborate the
potential of the proposed methods for real-world applications. Additional contributions
include: (i) a multimodal emotion database consisting of recordings of facial
expressions, body gestures, vocal expressions and physiological signals of actors enacting
various emotions, (ii) four multimodal deep belief network models and (iii)
an in-depth analysis of the effect of transfer of multimodal emotion features between
source and target networks on classification accuracy and training time. These related
contributions help comprehend the challenges involved in training deep learning
models and motivate the main goal of this dissertation.