• Encoding 3D contextual information for dynamic scene understanding 

  Hickson, Steven D. (Georgia Institute of Technology, 2020-04-27)

  This thesis aims to demonstrate how using 3D cues improves semantic labeling and object classification. Specifically, we will consider depth, surface normals, object classification, and pixel-wise semantic labeling in this ...

• Explaining model decisions and fixing them via human feedback 

  Ramasamy Selvaraju, Ramprasaath (Georgia Institute of Technology, 2020-05-07)

  Deep networks have enabled unprecedented breakthroughs in a variety of computer vision tasks. While these models enable superior performance, their increasing complexity and lack of decomposability into individually intuitive ...

• Domain adaptation via data augmentation 

  Gastineau, Eric Luc Adrien (Georgia Institute of Technology, 2020-04-28)

  Deep learning (DL) models require large labeled datasets for training. Practitioners often need to adapt an existing DL model to a different domain. For instance, a practitioner in a company developing autonomous vehicles ...

• Hidden fallacies in formally verified systems 

  Bobek, Jan (Georgia Institute of Technology, 2020-04-28)

  Formal verification or formal methods represent a rising trend in approaches to correct software construction, i.e. they help us answer the question of how to build software that contains no errors, colloquially known as ...

• Supporting healthy dyadic human relationships with power differentials using robots 

  Pettinati, Michael (Georgia Institute of Technology, 2020-04-23)

  Conflict is a natural part of ever-evolving human-human relationships. The way in which conflicts are handled can result in relationship growth, dissatisfaction (for one or both parties), or relationship dissolution. ...

• Policy-based exploration for efficient reinforcement learning 

  Subramanian, Kaushik (Georgia Institute of Technology, 2020-04-25)

  Reinforcement Learning (RL) is the field of research focused on solving sequential decision-making tasks modeled as Markov Decision Processes. Researchers have shown RL to be successful at solving a variety of problems ...

• 5-axis coverage path planning with deep reinforcement learning and fast parallel collision detection 

  Chen, Xin (Georgia Institute of Technology, 2020-04-24)

  5-axis machining is a strategy that allows computer numerical control (CNC) move an object or cutting tool along five different axes (X, Y, Z and two additional rotary axes) simultaneously. This provides infinite possibilities ...

• Disentangling neural network representations for improved generalization 

  Cogswell, Michael Andrew (Georgia Institute of Technology, 2020-04-24)

  Despite the increasingly broad perceptual capabilities of neural networks, applying them to new tasks requires significant engineering effort in data collection and model design. Generally, inductive biases can make this ...

• Convergence in min-max optimization 

  Lai, Kevin A. (Georgia Institute of Technology, 2020-04-20)

  Min-max optimization is a classic problem with applications in constrained optimization, robust optimization, and game theory. This dissertation covers new convergence rate results in min-max optimization. We show that the ...

• Hybrid and optical switching scheduling algorithms in data center networks 

  Liu, Liang (Georgia Institute of Technology, 2020-04-01)

  As a cost-effective approach to the data center network scalability problem, hybrid-switched data center networks have received considerable research attention recently. A hybrid-switched data center network employs a much ...

• Advanced machine learning approaches for characterization of transcriptional regulatory elements and genome-wide associations 

  Hassanzadeh, Hamid Reza (Georgia Institute of Technology, 2020-03-20)

  The deep learning revolution has initiated a surge of remarkable achievements in diverse research areas where large volumes of data that underlie complex processes exist. Despite the successful application of deep models ...

• Combatting abusive behavior in online communities using cross-community learning 

  Chandrasekharan, Eshwar (Georgia Institute of Technology, 2020-03-24)

  This dissertation aims to develop a deep understanding of abusive online behavior via statistical machine learning techniques to build tools that help counter it. I have developed computational approaches to model abusive ...

• Identifying opportunities to improve content moderation 

  Jhaver, Shagun (Georgia Institute of Technology, 2020-03-18)

  This thesis contributes a nuanced understanding of the challenges inherent in the design and implementation of fair and efficient content moderation systems. Using large-scale data analyses, participant observations, survey ...

• Human aspects of machine learning 

  Samadi, Samira (Georgia Institute of Technology, 2020-04-07)

  With the widespread use of Machine Learning (ML) algorithms in everyday life, it is important to study the human aspects of these algorithms. ML algorithms are increasingly used in applications that influence our day-to-day ...

• Building agents that can see, talk, and act 

  Das, Abhishek (Georgia Institute of Technology, 2020-04-25)

  A long-term goal in AI is to build general-purpose intelligent agents that simultaneously possess the ability to perceive the rich visual environment around us (through vision, audition, or other sensors), reason and infer ...

• Modeling for inversion in exploration geophysics 

  Louboutin, Mathias (Georgia Institute of Technology, 2020-03-16)

  Seismic inversion, and more generally geophysical exploration, aims at better understanding the earth's subsurface, which is one of today's most important challenges. Firstly, it contains natural resources that are critical ...

• Energy efficient architectures for irregular data streams 

  Srikanth, Sriseshan (Georgia Institute of Technology, 2020-03-27)

  An increasing prevalence of data-irregularity is being seen in applications today, particularly in machine learning, graph analytics, high-performance computing and cybersecurity. Faced with fundamental technology constraints, ...

• Manipulating state space distributions for sample-efficient imitation-learning 

  Schroecker, Yannick Karl Daniel (Georgia Institute of Technology, 2020-03-16)

  Imitation learning has emerged as one of the most effective approaches to train agents to act intelligently in unstructured and unknown domains. On its own or in combination with reinforcement learning, it enables agents ...

• Software and algorithms for large-scale seismic inverse problems 

  Witte, Philipp Andre (Georgia Institute of Technology, 2020-02-26)

  Seismic imaging and parameter estimation are an import class of inverse problems with practical relevance in resource exploration, carbon control and monitoring systems for geohazards. Seismic inverse problems involve ...

• Fault-tolerance on near-term quantum computers and subsystem quantum error correcting codes 

  Li, Muyuan (Georgia Institute of Technology, 2020-03-20)

  Large-scale fault-tolerant quantum computers are supposed to provide exponential speedup over many classical algorithms for solving realistic computationally intensive problems. Given that practical quantum computers are ...

View more