Divyam Anshumaan

Bio: Divyam Anshumaan is an academic researcher from Indraprastha Institute of Information Technology. The author has contributed to research in topics: Feature learning & Robustness (computer science). The author has an hindex of 1, co-authored 3 publications receiving 4 citations.

A Hybrid Approach to Tiger Re-Identification

Abstract: Visual data analytics is increasingly becoming an important part of wildlife monitoring and conservation strategies. In this work, we discuss our solution to the image-based Amur tiger re-identification (Re-ID) challenge hosted by the CVWC Workshop at ICCV 2019. Various factors like poor quality images, lighting and pose variations, and limited images per identity make tiger Re-ID a difficult task for deep learning models. Consequently, we propose to utilize both deep learning and traditional SIFT descriptor-based matching for tiger re-identification. The proposed deep network is based on a DenseNet model, fine-tuned by minimizing a classification cross-entropy loss regularized by a pairwise KL-divergence loss that promotes better semantically discriminative features. We also utilize several data transformations to improve the model's robustness and generalization across views and image quality variations. We establish the efficacy of our approach on the 'Plain Re-ID' challenge task by reporting results on the pre-cropped tiger Re-ID dataset. To further test our Re-ID model's robustness to detection quality, we also report results on the 'Wild Re-ID' task, which incorporates learning a tiger detection model. We show that our model is able to perform well on both the plain and wild Re-ID tasks. Code will be available at https://github.com/FGVC/DelPro.

WaveTransform: Crafting Adversarial Examples via Input Decomposition

Abstract: Frequency spectrum has played a significant role in learning unique and discriminating features for object recognition. Both low and high frequency information present in images have been extracted and learnt by a host of representation learning techniques, including deep learning. Inspired by this observation, we introduce a novel class of adversarial attacks, namely `WaveTransform', that creates adversarial noise corresponding to low-frequency and high-frequency subbands, separately (or in combination). The frequency subbands are analyzed using wavelet decomposition; the subbands are corrupted and then used to construct an adversarial example. Experiments are performed using multiple databases and CNN models to establish the effectiveness of the proposed WaveTransform attack and analyze the importance of a particular frequency component. The robustness of the proposed attack is also evaluated through its transferability and resiliency against a recent adversarial defense algorithm. Experiments show that the proposed attack is effective against the defense algorithm and is also transferable across CNNs.

WaveTransform: Crafting Adversarial Examples via Input Decomposition

Abstract: Frequency spectrum has played a significant role in learning unique and discriminating features for object recognition. Both low and high frequency information present in images have been extracted and learnt by a host of representation learning techniques, including deep learning. Inspired by this observation, we introduce a novel class of adversarial attacks, namely ‘WaveTransform’, that creates adversarial noise corresponding to low-frequency and high-frequency subbands, separately (or in combination). The frequency subbands are analyzed using wavelet decomposition; the subbands are corrupted and then used to construct an adversarial example. Experiments are performed using multiple databases and CNN models to establish the effectiveness of the proposed WaveTransform attack and analyze the importance of a particular frequency component. The robustness of the proposed attack is also evaluated through its transferability and resiliency against a recent adversarial defense algorithm. Experiments show that the proposed attack is effective against the defense algorithm and is also transferable across CNNs.

Simulating Network Paths with Recurrent Buffering Units

Abstract: Simulating physical network paths (e.g., Internet) is a cornerstone research problem in the emerging sub-field of AI-for-networking. We seek a model that generates end-to-end packet delay values in response to the time-varying load offered by a sender, which is typically a function of the previously output delays. The problem setting is unique, and renders the state-of-the-art text and time-series generative models inapplicable or ineffective. We formulate an ML problem at the intersection of dynamical systems, sequential decision making, and time-series modeling. We propose a novel grey-box approach to network simulation that embeds the semantics of physical network path in a new RNN-style model called Recurrent Buffering Unit, providing the interpretability of standard network simulator tools, the power of neural models, the efficiency of SGD-based techniques for learning, and yielding promising results on synthetic and real-world network traces.

ElephantBook: A Semi-Automated Human-in-the-Loop System for Elephant Re-Identification

Abstract: African elephants are vital to their ecosystems, but their populations are threatened by a rise in human-elephant conflict and poaching. Monitoring population dynamics is essential in conservation efforts; however, tracking elephants is a difficult task, usually relying on the invasive and sometimes dangerous placement of GPS collars. Although there have been many recent successes in the use of computer vision techniques for automated identification of other species, identification of elephants is extremely difficult and typically requires expertise as well as familiarity with elephants in the population. We have built and deployed a web-based platform and database for human-in-the-loop re-identification of elephants combining manual attribute labeling and state-of-the-art computer vision algorithms, known as ElephantBook. Our system is currently in use at the Mara Elephant Project, helping monitor the protected and at-risk population of elephants in the Greater Maasai Mara ecosystem. ElephantBook makes elephant re-identification usable by non-experts and scalable for use by multiple conservation NGOs.

How technology can advance the study of animal cognition in the wild

Abstract: Research on animal cognition has historically focused on the cognitive capabilities of a limited number of species through controlled laboratory studies. Animal cognition research is now broadening to study diverse taxa in their natural environments, which is facilitating insights into the fitness consequences of cognitive traits and a greater understanding of how ecological and social environments impact the expression of cognitive abilities. Many challenges remain regarding studying cognition in the wild, but recent technological advances have enabled researchers to overcome many of these challenges. Here, we highlight the technological advances leveraged for learning, social cognition, and spatial cognition field research, identify the current limitations of technology and barriers to its widespread use, and discuss promising technologies for field studies of animal cognition.

Exploring Robustness Connection between Artificial and Natural Adversarial Examples

Abstract: Although recent deep neural network algorithm has shown tremendous success in several computer vision tasks, their vulnerability against minute adversarial perturbations has raised a serious concern. In the early days of crafting these adversarial examples, artificial noises are optimized through the network and added in the images to decrease the confidence of the classifiers against the true class. However, recent efforts are showcasing the presence of natural adversarial examples which can also be effectively used to fool the deep neural networks with high confidence. In this paper, for the first time, we have raised the question that whether there is any robustness connection between artificial and natural adversarial examples. The possible robustness connection between natural and artificial adversarial examples is studied in the form that whether an adversarial example detector trained on artificial examples can detect the natural adversarial examples. We have analyzed several deep neural networks for the possible detection of artificial and natural adversarial examples in seen and unseen settings to set up a robust connection. The extensive experimental results reveal several interesting insights to defend the deep classifiers whether vulnerable against natural or artificially perturbed examples. We believe these findings can pave a way for the development of unified resiliency because defense against one attack is not sufficient for real-world use cases.