arXiv subjects

Dong-Yeon Cho

Publications and source records attributed to Dong-Yeon Cho.

6 recordsLinked to original sources

arXiv2025

Speech Tokenizer is Key to Consistent Representation

Speech tokenization is crucial in digital speech processing, converting continuous speech signals into discrete units for various computational tasks. This paper introduces a novel speech tokenizer with broad applicability across downstream tasks. While recent advances in residual vector quantization (RVQ) have incorporated semantic elements, they often neglect critical acoustic features. We propose an advanced approach that simultaneously encodes both linguistic and acoustic information, preserving prosodic and emotional content. Our method significantly enhances speech representation fidelity across diverse applications. Empirical evaluations demonstrate its effectiveness in speech coding, voice conversion, emotion recognition, and multimodal language modeling, without requiring additional training. This versatility underscores its potential as a key tool for advancing AI-driven speech processing.

cs.LG↗

arXiv2020

Domain-Agnostic Few-Shot Classification by Learning Disparate Modulators

Although few-shot learning research has advanced rapidly with the help of meta-learning, its practical usefulness is still limited because most of them assumed that all meta-training and meta-testing examples came from a single domain. We propose a simple but effective way for few-shot classification in which a task distribution spans multiple domains including ones never seen during meta-training. The key idea is to build a pool of models to cover this wide task distribution and learn to select the best one for a particular task through cross-domain meta-learning. All models in the pool share a base network while each model has a separate modulator to refine the base network in its own way. This framework allows the pool to have representational diversity without losing beneficial domain-invariant features. We verify the effectiveness of the proposed algorithm through experiments on various datasets across diverse domains.

cs.LG↗

arXiv2019

Discriminative Few-Shot Learning Based on Directional Statistics

Metric-based few-shot learning methods try to overcome the difficulty due to the lack of training examples by learning embedding to make comparison easy. We propose a novel algorithm to generate class representatives for few-shot classification tasks. As a probabilistic model for learned features of inputs, we consider a mixture of von Mises-Fisher distributions which is known to be more expressive than Gaussian in a high dimensional space. Then, from a discriminative classifier perspective, we get a better class representative considering inter-class correlation which has not been addressed by conventional few-shot learning algorithms. We apply our method to \emph{mini}ImageNet and \emph{tiered}ImageNet datasets, and show that the proposed approach outperforms other comparable methods in few-shot classification tasks.

cs.LG↗

arXiv2018

Auto-Meta: Automated Gradient Based Meta Learner Search

Fully automating machine learning pipelines is one of the key challenges of current artificial intelligence research, since practical machine learning often requires costly and time-consuming human-powered processes such as model design, algorithm development, and hyperparameter tuning. In this paper, we verify that automated architecture search synergizes with the effect of gradient-based meta learning. We adopt the progressive neural architecture search \cite{liu:pnas_google:DBLP:journals/corr/abs-1712-00559} to find optimal architectures for meta-learners. The gradient based meta-learner whose architecture was automatically found achieved state-of-the-art results on the 5-shot 5-way Mini-ImageNet classification problem with $74.65\%$ accuracy, which is $11.54\%$ improvement over the result obtained by the first gradient-based meta-learner called MAML \cite{finn:maml:DBLP:conf/icml/FinnAL17}. To our best knowledge, this work is the first successful neural architecture search implementation in the context of meta learning.

cs.LG↗

arXiv2018

Meta Continual Learning

Using neural networks in practical settings would benefit from the ability of the networks to learn new tasks throughout their lifetimes without forgetting the previous tasks. This ability is limited in the current deep neural networks by a problem called catastrophic forgetting, where training on new tasks tends to severely degrade performance on previous tasks. One way to lessen the impact of the forgetting problem is to constrain parameters that are important to previous tasks to stay close to the optimal parameters. Recently, multiple competitive approaches for computing the importance of the parameters with respect to the previous tasks have been presented. In this paper, we propose a learning to optimize algorithm for mitigating catastrophic forgetting. Instead of trying to formulate a new constraint function ourselves, we propose to train another neural network to predict parameter update steps that respect the importance of parameters to the previous tasks. In the proposed meta-training scheme, the update predictor is trained to minimize loss on a combination of current and past tasks. We show experimentally that the proposed approach works in the continual learning setting.

cs.LG↗

arXiv2017

NetREX: Network Rewiring using EXpression - Towards Context Specific Regulatory Networks

Understanding gene regulation is a fundamental step towards understanding of how cells function and respond to environmental cues and perturbations. An important step in this direction is to infer the transcription factor-gene regulatory network (GRN). However gene regulatory networks are typically constructed disregarding the fact that regulatory programs are conditioned on tissue type, developmental stage, sex, and other factors. Collecting multitude of features required for a reliable construction of GRNs such as physical features and functional features for every context of interest is costly. Therefore we need methods that is able to use the knowledge of a context-agnostic network for construction of a context specific regulatory network. To address this challenge we developed a computational approach that uses context specific expression data and a GRN constructed in a different but related context to construct a context specific GRN. Our method, NetREX, is inspired by network component analysis that estimates TF activities and their influences on target genes given predetermined topology of a TF-gene network. To predict a network under a different condition, NetREX removes the restriction that the topology of the TF-gene network is fixed and allows for adding and removing edges to that network. To solve the corresponding optimization problem, we provide a general mathematical framework allowing use of the recently proposed PALM technique and develop a convergent algorithm. We tested our NetREX on simulated data and subsequently applied it to gene expression data in adult females from Drosophila deletion (DrosDel) panel. The networks predicted by NetREX showed higher biological consistency than alternative approaches. In addition, we used the list of recently identified targets of the Doublesex (DSX) transcription factor to demonstrate the predictive power of our method.

q-bio.MN↗