Search Results for author: Mukund Sundararajan

Found 15 papers, 6 papers with code

Attributing AUC-ROC to Analyze Binary Classifier Performance

no code implementations24 May 2022 Arya Tafvizi, Besim Avci, Mukund Sundararajan

This observation leads to a simple, efficient attribution technique for examples (example attributions), and for pairs of examples (pair attributions).

First is Better Than Last for Language Data Influence

1 code implementation24 Feb 2022 Chih-Kuan Yeh, Ankur Taly, Mukund Sundararajan, Frederick Liu, Pradeep Ravikumar

However, we observe that since the activation connected to the last layer of weights contains "shared logic", the data influenced calculated via the last layer weights prone to a ``cancellation effect'', where the data influence of different examples have large magnitude that contradicts each other.

Reciprocity in Machine Learning

no code implementations19 Feb 2022 Mukund Sundararajan, Walid Krichene

Are these contributions (outflows of influence) and benefits (inflows of influence) reciprocal?

BIG-bench Machine Learning Recommendation Systems

The Penalty Imposed by Ablated Data Augmentation

no code implementations8 Jun 2020 Frederick Liu, Amir Najmi, Mukund Sundararajan

There is a set of data augmentation techniques that ablate parts of the input at random.

Data Augmentation

Attribution in Scale and Space

1 code implementation CVPR 2020 Shawn Xu, Subhashini Venugopalan, Mukund Sundararajan

Third, it eliminates the need for a 'baseline' parameter for Integrated Gradients [31] for perception tasks.

Object Recognition

Estimating Training Data Influence by Tracing Gradient Descent

3 code implementations NeurIPS 2020 Garima Pruthi, Frederick Liu, Mukund Sundararajan, Satyen Kale

We introduce a method called TracIn that computes the influence of a training example on a prediction made by the model.

The many Shapley values for model explanation

no code implementations ICML 2020 Mukund Sundararajan, Amir Najmi

In this paper, we use the axiomatic approach to study the differences between some of the many operationalizations of the Shapley value for attribution, and propose a technique called Baseline Shapley (BShap) that is backed by a proper uniqueness result.

Diabetes Prediction

How Important is a Neuron

no code implementations ICLR 2019 Kedar Dhamdhere, Mukund Sundararajan, Qiqi Yan

Informally, the conductance of a hidden unit of a deep network is the flow of attribution via this hidden unit.

A Note about: Local Explanation Methods for Deep Neural Networks lack Sensitivity to Parameter Values

no code implementations11 Jun 2018 Mukund Sundararajan, Ankur Taly

Local explanation methods, also known as attribution methods, attribute a deep network's prediction to its input (cf.

How Important Is a Neuron?

2 code implementations30 May 2018 Kedar Dhamdhere, Mukund Sundararajan, Qiqi Yan

Informally, the conductance of a hidden unit of a deep network is the \emph{flow} of attribution via this hidden unit.

Sentiment Analysis

Did the Model Understand the Question?

4 code implementations ACL 2018 Pramod Kaushik Mudrakarta, Ankur Taly, Mukund Sundararajan, Kedar Dhamdhere

Our strongest attacks drop the accuracy of a visual question answering model from $61. 1\%$ to $19\%$, and that of a tabular question answering model from $33. 5\%$ to $3. 3\%$.

Question Answering Visual Question Answering

It was the training data pruning too!

no code implementations12 Mar 2018 Pramod Kaushik Mudrakarta, Ankur Taly, Mukund Sundararajan, Kedar Dhamdhere

The large impact on the performance of the KDG model suggests that the pruning may be a useful pre-processing step in training other semantic parsers as well.

Question Answering

Axiomatic Attribution for Deep Networks

30 code implementations ICML 2017 Mukund Sundararajan, Ankur Taly, Qiqi Yan

We study the problem of attributing the prediction of a deep network to its input features, a problem previously studied by several other works.

Explainable artificial intelligence Interpretable Machine Learning

Gradients of Counterfactuals

no code implementations8 Nov 2016 Mukund Sundararajan, Ankur Taly, Qiqi Yan

Unfortunately, in nonlinear deep networks, not only individual neurons but also the whole network can saturate, and as a result an important input feature can have a tiny gradient.

Feature Importance Language Modelling +1

Cannot find the paper you are looking for? You can Submit a new open access paper.