BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//EE - ECPv5.10.0//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:EE
X-ORIGINAL-URL:https://ee.iisc.ac.in
X-WR-CALDESC:Events for EE
BEGIN:VTIMEZONE
TZID:Asia/Kolkata
BEGIN:STANDARD
TZOFFSETFROM:+0530
TZOFFSETTO:+0530
TZNAME:IST
DTSTART:20230101T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20230712T110000
DTEND;TZID=Asia/Kolkata:20230712T130000
DTSTAMP:20260525T235650
CREATED:20230711T065544Z
LAST-MODIFIED:20230711T065729Z
UID:240829-1689159600-1689166800@ee.iisc.ac.in
SUMMARY:[Oral Examination Talk] - A Learnable Distillation Approach For Model-agnostic Explainability With Multimodal Applications\, {Debarpan\, EE} [MMCR\, EE\, 11:00AM\, July 12]
DESCRIPTION:Title: A Learnable Distillation Approach For Model-agnostic Explainability With Multimodal Applications \nVenue: MMCR\, C241\, EE\, IISc\, and also in the Teams Link\n\n\nDate and Time: July 12\, 11:00AM.\n\n\n\nSpeaker: ​ Debarpan Bhattacharya\, MTech (Res) EE\, IISc\n\n\nAbstract: \nDeep neural networks are the most widely used examples of sophisticated mapping functions from feature space to class labels. In the recent years\, several high impact decisions in domains such as finance\, healthcare\, law and autonomous driving\, are made with deep models. In these tasks\, the model decisions lack interpretability\, and pose difficulties in making the models accountable. Hence\, there is a strong demand for developing explainable approaches which can elicit how the deep neural architecture generates the output decisions. \nThe current frameworks for explainability of model learning are based on gradients (eg. GradCAM\, guided-gradCAM\, Integrated gradients etc) or based on locally linear assumptions (eg. LIME). Some of these approaches require the knowledge of the deep model architecture\, which may be restrictive in many applications. Further\, most of the prior works in the literature highlight the results on a set of small number of examples to illustrate the performance of these XAI methods\, often lacking statistical evaluation. This talk proposes a new approach for explainability based on mask estimation approaches\, called the Distillation Approach for Model-agnostic Explainability (DAME). The DAME is a saliency-based explainability model that is post-hoc\, model-agnostic\, and applicable to any architecture/domain. The DAME is a student-teacher modeling approach\, where the teacher model is the original model for which the explainability is sought\, while the student model is the mask estimation model. The input sample is augmented with various data augmentation techniques to produce numerous samples in the immediate vicinity of the input. Using these samples\, the mask estimation model is learned to learn the saliency map of the input sample for predicting the labels. A distillation loss is used to train the DAME model\, and the student model tries to locally approximate the original model. Once the DAME model is trained\, the DAME generates a region of the input (either in space or in time-domain for images and audio samples\, respectively) that best explains the model predictions.  \nWe also propose an evaluation framework\, for both image and audio tasks\, where the XAI models are evaluated in a statistical framework on a set of held-out of examples with the Intersection-over-Union (IoU) metric. We have validated the DAME model for vision\, audio and biomedical tasks. Firstly\, we deploy the DAME for explaining a ResNet-50/ViT classifier pre-trained on ImageNet dataset for the object recognition task. Secondly\, we explain the predictions made by ResNet-50 classifier fine-tuned on Environmental Sound Classification (ESC-10) dataset for the audio event classification task. Finally\, we validate the DAME model on the COVID-19 classification task using cough audio recordings. In these tasks\, the DAME model is shown to outperform existing benchmarks for explainable modeling.  \n\n\n  \n​\n—————\n\nAll are welcome
URL:https://ee.iisc.ac.in/event/oral-examination-talk-a-learnable-distillation-approach-for-model-agnostic-explainability-with-multimodal-applications-debarpan-ee-mmcr-ee-1100am-july-12/
END:VEVENT
END:VCALENDAR