Ashwani Kumar

Adrien Deliège, Maxime Istasse, Ashwani Kumar et al.

In the framework of convolutional neural networks, downsampling is often performed with an average-pooling, where all the activations are treated equally, or with a max-pooling operation that only retains an element with maximum activation while discarding the others. Both of these operations are restrictive and have previously been shown to be sub-optimal. To address this issue, a novel pooling scheme, named\emph{ ordinal pooling}, is introduced in this work. Ordinal pooling rearranges all the elements of a pooling region in a sequence and assigns a different weight to each element based upon its order in the sequence. These weights are used to compute the pooling operation as a weighted sum of the rearranged elements of the pooling region. They are learned via a standard gradient-based training, allowing to learn a behavior anywhere in the spectrum of average-pooling to max-pooling in a differentiable manner. Our experiments suggest that it is advantageous for the networks to perform different types of pooling operations within a pooling layer and that a hybrid behavior between average- and max-pooling is often beneficial. More importantly, they also demonstrate that ordinal pooling leads to consistent improvements in the accuracy over average- or max-pooling operations while speeding up the training and alleviating the issue of the choice of the pooling operations and activation functions to be used in the networks. In particular, ordinal pooling mainly helps on lightweight or quantized deep learning architectures, as typically considered e.g. for embedded applications.

Ashwani Kumar

The thesis work presents assertion based functional verification of RTL representation of a digital design. The MBIST controller is designed based on a memory testing March algorithm. This March algorithm is a little modified March C algorithm which is modified by adding a paused element to test memory data retention faults. In assertion based functional verification, creation of verification plan, for MBIST controller RTL model and the implementation & simulation of the verification plan using System-Verilog and Synopsys-VCS are done. In ABV, verification plan includes the MBIST controller design and functional specification, functional coverage goals, code coverage goals, and assertions. Assertions are used to check the errors in RTL model of MBIST controller and to provide the functionality coverage. Functional coverage metrics are used to track the level or quality of verification. Most of the functional metrics score approximately reached the planned goal of 100 % which is planned in the verification plan. The designed MBIST controller is verified against the intended features. ABV approach helped to make the verification and design process efficient and less time-consuming by finding the bugs, exercising the corner cases in the design, and using the directed test cases in a small design. ABV helped to write directed and efficient test cases (25) which are approx 32 % less than the use of maximum possible random test cases (88) for designed MBIST controller with 100% assertion coverage and approximately equal total functional coverage, i.e., 97 % approx. In this way, ABV helped to fasten the design and verification process with better quality and assurance of correct functionality of MBIST controller after the integration in MBIST architecture.

Ashwani Kumar

In the framework of convolutional neural networks that lie at the heart of deep learning, downsampling is often performed with a max-pooling operation that only retains the element with maximum activation, while completely discarding the information contained in other elements in a pooling region. To address this issue, a novel pooling scheme, Ordinal Pooling Network (OPN), is introduced in this work. OPN rearranges all the elements of a pooling region in a sequence and assigns different weights to these elements based upon their orders in the sequence, where the weights are learned via the gradient-based optimisation. The results of our small-scale experiments on image classification task demonstrate that this scheme leads to a consistent improvement in the accuracy over max-pooling operation. This improvement is expected to increase in deeper networks, where several layers of pooling become necessary.