Deep Neural Networks (DNN) could forget the knowledge about earlier tasks when learning new tasks, and this is known as \textit{catastrophic forgetting}. While recent continual learning methods are capable of alleviating the catastrophic problem on toy-sized datasets, some issues still remain to be tackled when applying them in real-world problems. Recently, the fast mask-based learning method (e.g. piggyback \cite{mallya2018piggyback}) is proposed to address these issues by learning only a binary element-wise mask in a fast manner, while keeping the backbone model fixed. However, the binary mask has limited modeling capacity for new tasks. A more recent work \cite{hung2019compacting} proposes a compress-grow-based method (CPG) to achieve better accuracy for new tasks by partially training backbone model, but with order-higher training cost, which makes it infeasible to be deployed into popular state-of-the-art edge-/mobile-learning. The primary goal of this work is to simultaneously achieve fast and high-accuracy multi task adaption in continual learning setting. Thus motivated, we propose a new training method called \textit{kernel-wise Soft Mask} (KSM), which learns a kernel-wise hybrid binary and real-value soft mask for each task, while using the same backbone model. Such a soft mask can be viewed as a superposition of a binary mask and a properly scaled real-value tensor, which offers a richer representation capability without low-level kernel support to meet the objective of low hardware overhead. We validate KSM on multiple benchmark datasets against recent state-of-the-art methods (e.g. Piggyback, Packnet, CPG, etc.), which shows good improvement in both accuracy and training cost.

BENGALURU: At least 10 new AI-based jobs are being created for every 100 positions made redundant in traditional technology, experts in the field have said. However, the new roles are not being created as quickly as the positions that are eliminated and companies are increasingly training talent to fill the gap in artificial intelligence skills. This was the key takeaway from ET's recent conversation on Technology Talent Equilibrium with Sridhar Mitta, founder of digital technology services provider NextWealth Entrepreneurs; Prakash Mallya, the head of Intel India, and Supriyo Das, vice-president, Wipro Technologies. "If 100 jobs are cut, 10 jobs are created. The people who are losing jobs have different skills, and they're going away. People are getting new jobs on different skillsets," Mitta said.

The existence of an app to manage just about any financial operation today has led to an explosion of data. To cope with a rapid growth in information Banking, Financial Services and Insurance (BFSI) companies have been compelled to embrace a data-centric philosophy as part of their digital transformation journey. However, while embarking on this journey, companies should have a clear understanding of what they are trying to achieve or solve. This is what Bhavesh Lakhani, Senior Vice President and Head-IT, SBI Mutual Funds told Prakash Mallya, Vice President and Managing Director, Sales & Marketing Group of Intel India at the Intel Data-Centric Innovation Summit recently. Though SBI started its digital transformation journey three years ago intending to become the most preferred mutual fund company in India and overseas, physical forms continue to contribute over 80 percent of its revenues.

Reiterating its commitment to boost adoption of artificial intelligence (AI), Intel India today announced a developer community initiative – AI Developer Education Programme, aimed at educating 15,000 scientists, developers, analysts, and engineers. The educational programme is also aimed at deep learning and machine learning, the tech major said in a statement. The programme was announced at the first AI Day held in Bengaluru where thought-leaders from government, industry, and the academia congregated and discussed the potential of accelerating the AI revolution in the country. Under the programme, Intel will run 60 programmes across the year, ranging from workshops, roadshows, user group and senior technology leader round-tables. Announcing the programme, Intel South Asia managing director Prakash Mallya said data center and the intelligence behind the data collected can enable government and industry to make effective decisions based on algorithms.

Global chip maker Intel on Tuesday announced a string of initiatives to boost the usage of Artificial Intelligence (AI) in diverse sectors by collaborating with partners and customers across the country. "Our developer education programme will educate 15,000 scientists, developers, analysts and engineers on AI technologies, including Deep Learning and Machine Learning in India," said Intel South Asia Managing Director Praksh Mallya here. AI is a software programme that makes computers and machines think intelligently and faster with more predictability than a human mind. AI is also the main workload in data centres which operate in line with the Moore's Law of computing power doubling every year. By 2020, the industry expects more servers to process data analytics than other workloads and analytics predictors will be built into every application.

Artificial Intelligence (AI) is racing fast with booming start-ups, in terms of popularity and wide acceptability for its immense possibilities to boost productivity and propel economic growth. Every industry, be it e-commerce or education or fashion, is trying to introduce AI in every way conceivable to get the desired results. Prakash Mallya, Managing Director, Intel South Asia, shared his insight into Artificial Intelligence and talked about India's need for the same, on the sidelines of an event to mark Intel AI Day. "We conceptualized the AI Day a few months ago to bring together different industries, even communities and ecosystems to talk about Artificial Intelligence. Intel's point of view is that other than this one obvious reason, the event can also serve as an opportunity to seek and receive feedback on the same, hold discussions with ecosystems on how we can work jointly to accelerate the implementation or usage of AI and transform lives of every individual in the country," he said.

Intel India announced the launch of its Artificial Intelligence Developer Program aimed at educating 15,000 individuals - scientists, developers, analysts and engineers - on AI technologies in a bid to boost the adoption of Artificial Intelligence (AI) in the homeland. Aimed at promoting AI amongst Indian developers, Intel will organise 60 programs across the year, ranging from lectures, workshops, roadshows and roundtables. The initiative aims at imparting knowledge on the basics of adoption of AI via technical sessions. As India undergoes digital transformation, the intelligent analysis of the data collected will enable the government and the industry to make effective decisions offering increasing opportunities for using AI in the country. "To make this happen, Intel is collaborating with Hewlett Packard Enterprise, Wipro, Julia Computing and Calligo Technologies, by enabling them with AI solutions based on Intel architecture," said Prakash Mallya, Managing Director, Intel South Asia.

Global chip maker Intel on Tuesday announced a string of initiatives to boost the usage of Artificial Intelligence (AI) in diverse sectors by collaborating with partners and customers across the country. "Our developer education programme will educate 15,000 scientists, developers, analysts and engineers on AI technologies, including Deep Learning and Machine Learning in India," said Intel South Asia Managing Director Praksh Mallya here. AI is a software programme that makes computers and machines think intelligently and faster with more predictability than a human mind. AI is also the main workload in data centres which operate in line with the Moore's Law of computing power doubling every year. By 2020, the industry expects more servers to process data analytics than other workloads and analytics predictors will be built into every application.