AIConfPaper
← Search

Mayank Goswami

8 accepted papers

2026

The benefits of full data shuffle, now with optimal I/O cost: $k$-wise independence and matrix transposition to the rescue

Peyman Afshani, Rezaul Chowdhury, Mayank Goswami, Jens Kristian R Schou, Francesco Silvestri, Mariafiore Tognon
ICML 2026poster

It is known that RandomShuffle, the without replacement version of Stochastic Gradient Descend (SGD), converges faster than with-replacement SGD. However, RandomShuffle requires to uniformly perform a random permutation of the input sequence, which is known to have an high I/O complexity due to data…

Cited by 0Source ↗Scholar ↗
2023

Learning to Segment from Noisy Annotations: A Spatial Correction Approach

Jiachen Yao, Yikai Zhang, Songzhu Zheng, Mayank Goswami, Prateek Prasanna, Chao Chen
ICLR 2023poster

Noisy labels can significantly affect the performance of deep neural networks (DNNs). In medical image segmentation tasks, annotations are error-prone due to the high demand in annotation time and in the annotators' expertise. Existing methods mostly tackle label noise in classification tasks. Their…

Cited by 18Source ↗PDF ↗code ↗Scholar ↗
2022

A Manifold View of Adversarial Risk

Wenjia Zhang, Yikai Zhang, Xiaoling Hu, Mayank Goswami, Chao Chen, Dimitris N. Metaxas
AISTATS 2022poster

The adversarial risk of a machine learning model has been widely studied. Most previous works assume that the data lies in the whole ambient space. We propose to take a new angle and take the manifold assumption into consideration. Assuming data lies in a manifold, we investigate two new types of ad…

Cited by 4Source ↗PDF ↗Scholar ↗
2022

Stability of SGD: Tightness analysis and improved bounds

Yikai Zhang, Wenjia Zhang, Sammy Bald, Vamsi Pingali, Chao Chen, Mayank Goswami
UAI 2022poster

Stochastic Gradient Descent (SGD) based methods have been widely used for training large-scale machine learning models that also generalize well in practice. Several explanations have been offered for this generalization performance, a prominent one being algorithmic stability Hardt et al [2016]. Ho…

Cited by 41Source ↗PDF ↗Scholar ↗
2021

Learning with Feature-Dependent Label Noise: A Progressive Approach

Yikai Zhang, Songzhu Zheng, Pengxiang Wu, Mayank Goswami, Chao Chen
ICLR 2021spotlight

Label noise is frequently observed in real-world large-scale datasets. The noise is introduced due to a variety of reasons; it is heterogeneous and feature-dependent. Most existing approaches to handling noisy labels fall into two categories: they either assume an ideal feature-independent noise, or…

Cited by 195Source ↗PDF ↗code ↗Scholar ↗
2021

Topological Detection of Trojaned Neural Networks

Songzhu Zheng, Yikai Zhang, Hubert Wagner, Mayank Goswami, Chao Chen
NeurIPS 2021poster

Deep neural networks are known to have security issues. One particular threat is the Trojan attack. It occurs when the attackers stealthily manipulate the model's behavior through Trojaned training samples, which can later be exploited. Guided by basic neuroscientific principles, we discover subtle…

Cited by 58Source ↗PDF ↗Scholar ↗
2020

A Topological Filter for Learning with Label Noise

Pengxiang Wu, Songzhu Zheng, Mayank Goswami, Dimitris Metaxas, Chao Chen
NeurIPS 2020poster

Noisy labels can impair the performance of deep neural networks. To tackle this problem, in this paper, we propose a new method for filtering label noise. Unlike most existing methods relying on the posterior probability of a noisy classifier, we focus on the much richer spatial behavior of data in…

Cited by 145Source ↗PDF ↗code ↗Scholar ↗
2020

Error-Bounded Correction of Noisy Labels

Songzhu Zheng, Pengxiang Wu, Aman Goswami, Mayank Goswami, Dimitris Metaxas, Chao Chen
ICML 2020poster

To collect large scale annotated data, it is inevitable to introduce label noise, i.e., incorrect class labels. To be robust against label noise, many successful methods rely on the noisy classifiers (i.e., models trained on the noisy training data) to determine whether a label is trustworthy. Howev…

Cited by 171Source ↗PDF ↗code ↗Scholar ↗