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 41 for “"Disentanglement"”.
-
Sparse Expansion and Neuronal Disentanglement
We show how to improve the inference efficiency of an LLM by expanding it into a mixture of sparse experts, where each expert is a copy of the same weights and one-shot pruned for a specific cluster of input values. We call this approach Sparse Expansion. We show that for models like Llama 2 7B, as …
-
Capsule Networks: Framework and Application to Disentanglement for Generative Models
… networks. Thus, we attempt to increase the disentanglement of latent variables in variational autoencoders without compromising the generated image quality. In this thesis, a novel generative model based on capsule networks and a variational autoencoder is proposed. Motivated by the concept …
-
Interpreting Deep Neural Networks and Beyond: Visualization, Learning Dynamics, and Disentanglement
… visualization, learning dynamics, and disentanglement. First, the demand for human interpretable explanations for model decisions has driven the development of visualization techniques. In particular, a class of backpropagation-based visualizations has attracted much attention recently, …
-
SpeechSplit2: An efficient unsupervised speech disentanglement model for multi-aspect voice conversion
Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2024-12-01
-
Deep Generative Models for Unsupervised Scale-Based and Position-Based Disentanglement of Concepts from Face Images.
Among the different categories of natural images, face images are very important because of the role they play in human social interactions. It is recognised that despite all the recent advances of artificial intelligence using deep neural networks, computers are still struggling at achieving a …
-
Disentangling neural network representations for improved generalization
… In this thesis we identify three kinds of disentanglement, implement a strategy for enforcing disentanglement in each case, and show that more general representations result. These perspectives treat disentanglement as statistical independence of features in image classification, language …
-
Direct imaging of single-walled carbon nanotubes reptational dynamics in packed-colloid porous media
… pore heterogeneity causes wide distributed SWCNT disentanglement time between pores, which causes SWCNT location dependent Brownian rotation dynamics from subdiffusive to diffusive regime. By studying the orientation and SWCNT pore disentanglement in packed-colloid pore, we study how the …
-
The Use of Group Art Therapy to Facilitate the Development of Identity with Adolescents in a Residential Substance Abuse Treatment Center
… review focuses on three main areas: the disentanglement of the sober identity, adolescent development, and the use of art as an intervention within the realms of substance abuse. The recovery aspect of the intervention has a theoretical foundation in Goffman’s (1963) concept of …
-
Numerical Methods in Deep Learning and Computer Vision
… We first investigate the relation between disentanglement and orthogonality -- the generative models are enforced with different proposed orthogonality to show that the disentanglement performance is indeed improved. We also challenge the linear assumption of the latent traversal paths and …
-
The process of coping with dating violence: a qualitative study
… couple imbalances, seductive processes, disentanglement processes and paradigmatic shifts. In essence, it was discovered that vulnerable young women who formed fused, imbalanced relationships with vulnerable men tended to use system-maintaining coping strategies to deal with the violence …
-
Effects of added drag on cetaceans : fishing gear entanglement and external tag attachment
… drag and buoyancy from fishing gear and through disentanglement: some swimming patterns were consistently modulated in response. Disentanglement significantly altered dive behavior, and can affect thrust production. Changes in the force balance and swimming behaviors have implications for the …
-
On Disentangled Analysis-by-Synthesis Shape Representations
… reconstructive signal, and the opportunity for disentanglement, via regularizing priors. The resulting representations are more versatile, controllable, and widely applicable. We begin by defining a prior on the deformation space of shapes, separating intrinsic versus isometric changes through …
-
Effects and added drag on cetaceans : fishing gear entanglement and external tag
… drag and buoyancy from fishing gear and through disentanglement: some swimming patterns were consistently modulated in response. Disentanglement significantly altered dive behavior, and can affect thrust production. Changes in the force balance and swimming behaviors have implications for the …
-
Learning Low-Level Priors from Images for Inference and Synthesis
… facilitate inference of geometry, static-dynamic disentanglement, and material properties. We present a self-supervised method to construct a persistent representation for inferring geometry and appearance inferred using a single image at test time. This representation can be leveraged to infer …
-
Durability study of proton exchange membrane fuel cells via experimental investigations and mathematical modeling
… before yielding and polymer aggregates disentanglement/ reorientation after yielding. In addition, the influence of uniaxial loading on proton conductivity of N117 and BPSH35 membranes is investigated. When the membranes are stretched, their proton conductivities in the straining …
-
Disentangling time constant and time dependent hidden state in time series with variational Bayesian inference
… datasets. This approach demonstrates effective disentanglement, motivating future work to explore the efficacy of this mo del in real word datasets.
-
Towards Data Efficiency and Controllable Representations for Deep Learning in Resource-Constrained Domains
… evaluation metrics to assess the quality of disentanglement remain underdeveloped, limiting their practical utility. These limitations highlight the need for advancements in data-efficient machine learning and controllable generative modelling, focusing on domain-specific validity and …
-
Photoelectrons and Recombining Electrons in Atomic and Molecular Systems
… rescattering dynamics. In Chapters 5-6, the disentanglement of the valence electronic structure, as well as evidence for low-energy shape resonances, of the pyrimidine-type nucleobases, thymine, uracil, and cytosine, is shown using high-resolution photoelectron spectroscopy. Chapters 7-8 …
-
Towards Comprehensive Visual Understanding via Deep Neural Networks
… ii)), this thesis introduces a holistic disentanglement and modeling for segment targets under an identical framework. Extensive experiments are conducted to verify the effectiveness of the proposed methods on scene understanding tasks, including Unsupervised Domain Adaptation (UDA), …
-
Developing Better Models for Dialogue Threads and Responses
… history by proposing novel models for dialogue disentanglement. Specifically, multiple threads of discussions in the same dialogue are disentangled and organized to single topic threads. This research can help users to more easily follow up on the topics discussed in complex dialogues. We …
Page 1 of 3