1 code implementation • 21 Feb 2024 • Nathan Beck, Adithya Iyer, Rishabh Iyer
As supervised fine-tuning of pre-trained models within NLP applications increases in popularity, larger corpora of annotated data are required, especially with increasing parameter counts in large language models.
no code implementations • 21 Feb 2024 • Nathan Beck, Truong Pham, Rishabh Iyer
With increasing volume of data being used across machine learning tasks, the capability to target specific subsets of data becomes more important.
1 code implementation • 13 Jan 2024 • Durga Sivasubramanian, Lokesh Nagalapatti, Rishabh Iyer, Ganesh Ramakrishnan
We conduct experiments using four real-world datasets and show that GCFL is (1) more compute and energy efficient than FL, (2) robust to various kinds of noise in both the feature space and labels, (3) preserves the privacy of the validation dataset, and (4) introduces a small communication overhead but achieves significant gains in performance, particularly in cases when the clients' data is noisy.
no code implementations • 29 Sep 2023 • Anay Majee, Suraj Kothawade, Krishnateja Killiamsetty, Rishabh Iyer
In this paper, we introduce the SCoRe (Submodular Combinatorial Representation Learning) framework and propose a family of Submodular Combinatorial Loss functions to overcome these pitfalls in contrastive learning.
no code implementations • 2 Jun 2023 • Nathan Beck, KrishnaTeja Killamsetty, Suraj Kothawade, Rishabh Iyer
Active Learning (AL) is a human-in-the-loop framework to interactively and adaptively label data instances, thereby enabling significant gains in model performance compared to random sampling.
1 code implementation • 18 May 2023 • Nathan Beck, Suraj Kothawade, Pradeep Shenoy, Rishabh Iyer
However, learning unbiased models depends on building a dataset that is representative of a diverse range of realistic scenarios for a given task.
no code implementations • 11 May 2023 • H S V N S Kowndinya Renduchintala, KrishnaTeja Killamsetty, Sumit Bhatia, Milan Aggarwal, Ganesh Ramakrishnan, Rishabh Iyer, Balaji Krishnamurthy
A salient characteristic of pre-trained language models (PTLMs) is a remarkable improvement in their generalization capability and emergence of new capabilities with increasing model capacity and pre-training dataset size.
no code implementations • 30 Jan 2023 • KrishnaTeja Killamsetty, Alexandre V. Evfimievski, Tejaswini Pedapati, Kiran Kate, Lucian Popa, Rishabh Iyer
Training deep networks and tuning hyperparameters on large datasets is computationally intensive.
no code implementations • 30 Oct 2022 • Ashish Mittal, Durga Sivasubramanian, Rishabh Iyer, Preethi Jyothi, Ganesh Ramakrishnan
Training state-of-the-art ASR systems such as RNN-T often has a high associated financial and environmental cost.
no code implementations • 4 Oct 2022 • Suraj Kothawade, Atharv Savarkar, Venkat Iyer, Lakshman Tamil, Ganesh Ramakrishnan, Rishabh Iyer
It is often the case that a suboptimal performance is obtained on some classes due to the natural class imbalance issue that comes with medical data.
no code implementations • 4 Oct 2022 • Suraj Kothawade, Akshit Srivastava, Venkat Iyer, Ganesh Ramakrishnan, Rishabh Iyer
Avoiding out-of-distribution (OOD) data is critical for training supervised machine learning models in the medical imaging domain.
no code implementations • 17 Jun 2022 • Suraj Kothawade, Shivang Chopra, Saikat Ghosh, Rishabh Iyer
Most approaches assume access to a seed set of instances which contain these rare data instances.
1 code implementation • 15 Mar 2022 • KrishnaTeja Killamsetty, Guttu Sai Abhishek, Aakriti, Alexandre V. Evfimievski, Lucian Popa, Ganesh Ramakrishnan, Rishabh Iyer
Our central insight is that using an informative subset of the dataset for model training runs involved in hyper-parameter optimization, allows us to find the optimal hyper-parameter configuration significantly faster.
no code implementations • 10 Mar 2022 • Suraj Kothawade, Pavan Kumar Reddy, Ganesh Ramakrishnan, Rishabh Iyer
This issue is further pronounced in SSL methods, as they would use this biased model to obtain psuedo-labels (on the unlabeled data) during training.
1 code implementation • 22 Feb 2022 • Vishal Kaushal, Ganesh Ramakrishnan, Rishabh Iyer
A recent work has also leveraged submodular functions to propose submodular information measures which have been found to be very useful in solving the problems of guided subset selection and guided summarization.
1 code implementation • 30 Jan 2022 • Changbin Li, Suraj Kothawade, Feng Chen, Rishabh Iyer
Meta learning has proven to be able to learn a parametrized model for FSC by training on various other classification tasks.
no code implementations • NeurIPS 2021 • Ping Zhang, Rishabh Iyer, Ashish Tendulkar, Gaurav Aggarwal, Abir De
Marked temporal point processes (MTPPs) have emerged as a powerful modelingtool for a wide variety of applications which are characterized using discreteevents localized in continuous time.
1 code implementation • 30 Nov 2021 • Suraj Kothawade, Saikat Ghosh, Sumit Shekhar, Yu Xiang, Rishabh Iyer
We propose TALISMAN, a novel framework for Targeted Active Learning or object detectIon with rare slices using Submodular MutuAl iNformation.
no code implementations • CVPR 2022 • Rishabh Tiwari, KrishnaTeja Killamsetty, Rishabh Iyer, Pradeep Shenoy
To address this, replay-based CL approaches maintain and repeatedly retrain on a small buffer of data selected across encountered tasks.
no code implementations • 10 Oct 2021 • Suraj Kothawade, Anmol Mekala, Chandra Sekhara D, Mayank Kothyari, Rishabh Iyer, Ganesh Ramakrishnan, Preethi Jyothi
To address this problem, we propose DITTO (Data-efficient and faIr Targeted subseT selectiOn) that uses Submodular Mutual Information (SMI) functions as acquisition functions to find the most informative set of utterances matching a target accent within a fixed budget.
Automatic Speech Recognition Automatic Speech Recognition (ASR) +1
1 code implementation • Findings (ACL) 2022 • Ayush Maheshwari, KrishnaTeja Killamsetty, Ganesh Ramakrishnan, Rishabh Iyer, Marina Danilevsky, Lucian Popa
These LFs, in turn, have been used to generate a large amount of additional noisy labeled data, in a paradigm that is now commonly referred to as data programming.
1 code implementation • 1 Aug 2021 • Guttu Sai Abhishek, Harshad Ingole, Parth Laturia, Vineeth Dorna, Ayush Maheshwari, Rishabh Iyer, Ganesh Ramakrishnan
SPEAR facilitates weak supervision in the form of heuristics (or rules) and association of noisy labels to the training dataset.
1 code implementation • NeurIPS 2021 • Suraj Kothawade, Nathan Beck, KrishnaTeja Killamsetty, Rishabh Iyer
Active learning has proven to be useful for minimizing labeling costs by selecting the most informative samples.
1 code implementation • 23 Jun 2021 • Durga Sivasubramanian, Rishabh Iyer, Ganesh Ramakrishnan, Abir De
First, we represent this problem with simplified constraints using the dual of the original training problem and show that the objective of this new representation is a monotone and alpha-submodular function, for a wide variety of modeling choices.
no code implementations • 16 Jun 2021 • Nathan Beck, Durga Sivasubramanian, Apurva Dani, Ganesh Ramakrishnan, Rishabh Iyer
Issues in the current literature include sometimes contradictory observations on the performance of different AL algorithms, unintended exclusion of important generalization approaches such as data augmentation and SGD for optimization, a lack of study of evaluation facets like the labeling efficiency of AL, and little or no clarity on the scenarios in which AL outperforms random sampling (RS).
1 code implementation • NeurIPS 2021 • KrishnaTeja Killamsetty, Xujiang Zhao, Feng Chen, Rishabh Iyer
In this work, we propose RETRIEVE, a coreset selection framework for efficient and robust semi-supervised learning.
1 code implementation • 3 Jun 2021 • Mustafa Safa Ozdayi, Murat Kantarcioglu, Rishabh Iyer
In such a setting, we design fair training algorithms which exhibit both good utility, and low bias.
1 code implementation • Findings (ACL) 2021 • Atul Sahay, Anshul Nasery, Ayush Maheshwari, Ganesh Ramakrishnan, Rishabh Iyer
We introduce a novel formulation that takes advantage of the syntactic grammar rules and is independent of the base system.
no code implementations • 30 Apr 2021 • Suraj Kothawade, Vishal Kaushal, Ganesh Ramakrishnan, Jeff Bilmes, Rishabh Iyer
With the rapid growth of data, it is becoming increasingly difficult to train or improve deep learning models with the right subset of data.
1 code implementation • 27 Feb 2021 • Suraj Kothawade, Vishal Kaushal, Ganesh Ramakrishnan, Jeff Bilmes, Rishabh Iyer
Examples of such problems include: i)targeted learning, where the goal is to find subsets with rare classes or rare attributes on which the model is underperforming, and ii)guided summarization, where data (e. g., image collection, text, document or video) is summarized for quicker human consumption with specific additional user intent.
3 code implementations • 27 Feb 2021 • KrishnaTeja Killamsetty, Durga Sivasubramanian, Ganesh Ramakrishnan, Abir De, Rishabh Iyer
We show rigorous theoretical and convergence guarantees of the proposed algorithm and, through our extensive experiments on real-world datasets, show the effectiveness of our proposed framework.
no code implementations • 26 Jan 2021 • Vishal Kaushal, Suraj Kothawade, Anshul Tomar, Rishabh Iyer, Ganesh Ramakrishnan
For long videos, human reference summaries necessary for supervised video summarization techniques are difficult to obtain.
1 code implementation • 19 Dec 2020 • KrishnaTeja Killamsetty, Durga Sivasubramanian, Ganesh Ramakrishnan, Rishabh Iyer
Finally, we propose Glister-Active, an extension to batch active learning, and we empirically demonstrate the performance of Glister on a wide range of tasks including, (a) data selection to reduce training time, (b) robust learning under label noise and imbalance settings, and (c) batch-active learning with several deep and shallow models.
no code implementations • 13 Nov 2020 • KrishnaTeja Killamsetty, Changbin Li, Chen Zhao, Rishabh Iyer, Feng Chen
Model-Agnostic Meta-Learning (MAML), a popular gradient-based meta-learning framework, assumes that the contribution of each task or instance to the meta-learner is equal.
no code implementations • 16 Oct 2020 • Suraj Kothawade, Jiten Girdhar, Chandrashekhar Lavania, Rishabh Iyer
Unfortunately, these models only learn the relative importance of the different submodular functions (such as diversity, representation or importance), but cannot learn more complex feature representations, which are often required for state-of-the-art performance.
no code implementations • 14 Oct 2020 • Mustafa Safa Ozdayi, Murat Kantarcioglu, Rishabh Iyer
Particularly, in settings where local data distributions vastly differ among agents, FL performs rather poorly with respect to the centralized training.
no code implementations • 12 Oct 2020 • Vishal Kaushal, Suraj Kothawade, Ganesh Ramakrishnan, Jeff Bilmes, Himanshu Asnani, Rishabh Iyer
We study submodular information measures as a rich framework for generic, query-focused, privacy sensitive, and update summarization tasks.
1 code implementation • 7 Oct 2020 • Xujiang Zhao, Killamsetty Krishnateja, Rishabh Iyer, Feng Chen
This work addresses the following question: How do out-of-distribution (OOD) data adversely affect semi-supervised learning algorithms?
1 code implementation • Findings (ACL) 2021 • Ayush Maheshwari, Oishik Chatterjee, KrishnaTeja Killamsetty, Ganesh Ramakrishnan, Rishabh Iyer
The first contribution of this work is an introduction of a framework, \model which is a semi-supervised data programming paradigm that learns a \emph{joint model} that effectively uses the rules/labelling functions along with semi-supervised loss functions on the feature space.
no code implementations • 29 Jul 2020 • Vishal Kaushal, Suraj Kothawade, Rishabh Iyer, Ganesh Ramakrishnan
Thirdly, we demonstrate that in the presence of multiple ground truth summaries (due to the highly subjective nature of the task), learning from a single combined ground truth summary using a single loss function is not a good idea.
no code implementations • 27 Jun 2020 • Rishabh Iyer, Jeff Bilmes
In this paper, we try to provide a more complete picture of the relationship between submodularity with concavity.
no code implementations • 27 Jun 2020 • Rishabh Iyer, Ninad Khargonkar, Jeff Bilmes, Himanshu Asnani
In this paper, we study combinatorial information measures that generalize independence, (conditional) entropy, (conditional) mutual information, and total correlation defined over sets of (not necessarily random) variables.
no code implementations • 25 Jan 2020 • Rishabh Iyer
While several existing papers have studied robust submodular maximization, ours is the first work to study the minimization version under a broad range of combinatorial constraints including cardinality, knapsack, matroid as well as graph-based constraints such as cuts, paths, matchings, and trees.
no code implementations • 14 Jun 2019 • Rishabh Iyer
Furthermore, we also study robust submodular minimization and maximization under multiple submodular upper and lower bound constraints.
no code implementations • 26 Feb 2019 • Rishabh Iyer, Jeff Bilmes
In this paper, we investigate a class of submodular problems which in general are very hard.
no code implementations • 26 Feb 2019 • Rishabh Iyer, Jeff Bilmes
We are motivated by large scale submodular optimization problems, where standard algorithms that treat the submodular functions in the \emph{value oracle model} do not scale.
no code implementations • 3 Jan 2019 • Vishal Kaushal, Rishabh Iyer, Khoshrav Doctor, Anurag Sahoo, Pratik Dubal, Suraj Kothawade, Rohan Mahadev, Kunal Dargan, Ganesh Ramakrishnan
This paper addresses automatic summarization of videos in a unified manner.
1 code implementation • 3 Jan 2019 • Vishal Kaushal, Rishabh Iyer, Suraj Kothawade, Rohan Mahadev, Khoshrav Doctor, Ganesh Ramakrishnan
Supervised machine learning based state-of-the-art computer vision techniques are in general data hungry.
no code implementations • 24 Sep 2018 • Vishal Kaushal, Sandeep Subramanian, Suraj Kothawade, Rishabh Iyer, Ganesh Ramakrishnan
We propose a novel framework for domain specific video summarization.
1 code implementation • 24 Sep 2018 • Rishabh Iyer, Pratik Dubal, Kunal Dargan, Suraj Kothawade, Rohan Mahadev, Vishal Kaushal
With increasing amounts of visual data being created in the form of videos and images, visual data selection and summarization are becoming ever increasing problems.
1 code implementation • 12 Sep 2018 • Rishabh Iyer, Nimit Acharya, Tanuja Bompada, Denis Charles, Eren Manavoglu
Through extensive experiments, we demonstrate the utility of of our OL framework; how the two OL schemes relate to each other and how they trade-off between the new and historical data.
2 code implementations • 17 Jul 2018 • Rishabh Iyer, John T. Halloran, Kai Wei
This paper introduces Jensen, an easily extensible and scalable toolkit for production-level machine learning and convex optimization.
no code implementations • 28 May 2018 • Vishal Kaushal, Anurag Sahoo, Khoshrav Doctor, Narasimha Raju, Suyash Shetty, Pankaj Singh, Rishabh Iyer, Ganesh Ramakrishnan
Supervised machine learning based state-of-the-art computer vision techniques are in general data hungry and pose the challenges of not having adequate computing resources and of high costs involved in human labeling efforts.
no code implementations • 27 May 2018 • Pratik Dubal, Rohan Mahadev, Suraj Kothawade, Kunal Dargan, Rishabh Iyer
To our knowledge, this is the first work which provides a comprehensive evaluation of different deep learning models on various real-world customer deployment scenarios of surveillance video analytics.
no code implementations • 18 Apr 2018 • John Moore, Joel Pfeiffer, Kai Wei, Rishabh Iyer, Denis Charles, Ran Gilad-Bachrach, Levi Boyles, Eren Manavoglu
In real world systems, the predictions of deployed Machine Learned models affect the training data available to build subsequent models.
no code implementations • 4 Apr 2017 • Anurag Sahoo, Vishal Kaushal, Khoshrav Doctor, Suyash Shetty, Rishabh Iyer, Ganesh Ramakrishnan
Most importantly, we also show that we can summarize hours of video data in a few seconds, and our system allows the user to generate summaries of various lengths and types interactively on the fly.
no code implementations • NeurIPS 2015 • Jennifer Gillenwater, Rishabh Iyer, Bethany Lusch, Rahul Kidambi, Jeff Bilmes
We show that there is a largely unexplored class of functions (positive polymatroids) that can define proper discrete metrics over pairs of binary vectors and that are fairly tractable to optimize over.
no code implementations • NeurIPS 2015 • Kai Wei, Rishabh Iyer, Shengjie Wang, Wenruo Bai, Jeff Bilmes
While the robust versions have been studied in the theory community, existing work has focused on tight approximation guarantees, and the resultant algorithms are not, in general, scalable to very large real-world applications.
no code implementations • 24 Jun 2015 • Rishabh Iyer, Jeff Bilmes
This manuscript provides a more complete picture on the relationship between submodularity with convexity and concavity, by extending many of the results connecting submodularity with convexity to the concave aspects of submodularity.
Discrete Mathematics Data Structures and Algorithms
no code implementations • 9 Aug 2014 • Rishabh Iyer, Jeff A. Bilmes
We show how a number of recently used web ranking models are forms of Lovasz-Bregman rank aggregation and also observe that a natural form of Mallow's model using the LB divergence has been used as conditional ranking models for the "Learning to Rank" problem.
no code implementations • 9 Aug 2014 • Rishabh Iyer, Jeff A. Bilmes
We extend the work of Narasimhan and Bilmes [30] for minimizing set functions representable as a dierence between submodular functions.
no code implementations • NeurIPS 2013 • Rishabh Iyer, Jeff Bilmes
We are motivated by a number of real-world applications in machine learning including sensor placement and data subset selection, which require maximizing a certain submodular function (like coverage or diversity) while simultaneously minimizing another (like cooperative cost).
no code implementations • NeurIPS 2013 • Rishabh Iyer, Stefanie Jegelka, Jeff Bilmes
We either use a black-box transformation of the function (for approximation and learning), or a transformation of algorithms to use an appropriate surrogate function (for minimization).
no code implementations • 24 Aug 2013 • Rishabh Iyer, Jeff Bilmes
We show how a number of recently used web ranking models are forms of Lovasz-Bregman rank aggregation and also observe that a natural form of Mallow's model using the LB divergence has been used as conditional ranking models for the 'Learning to Rank' problem.
no code implementations • 5 Aug 2013 • Rishabh Iyer, Stefanie Jegelka, Jeff Bilmes
We present a practical and powerful new framework for both unconstrained and constrained submodular function optimization based on discrete semidifferentials (sub- and super-differentials).
no code implementations • 3 Jul 2012 • Rishabh Iyer, Jeff Bilmes
We extend the work of Narasimhan and Bilmes [30] for minimizing set functions representable as a difference between submodular functions.