The wide proliferation of internet and communication technologies since mid-1990s has automated information sharing to an unimaginable extent.

"Cyber Landscape" is defined as the physical, logical, computational, communication and control components of ICT infrastructure. Threats and vulnerabilities are defined by capability levels to withstand and face hostility.

ICT components like Software, Hardware, communication, computation, protocol etchave been the root nodes of attack. The location of attacksareat global, national, social, organizational, institutional and at all levels of administrative levels of governance.

ICT infrastructure carry, process and control moment to moment business, financial, economical, banking, commerce, scientific, engineering, social, industrial, cultural, governmental, classified, educational and personal communications between countries and within a country.

Threats, vulnerabilities and risks to information sharing and systems built on ICT seldom change. But the level of sophistication of cyber threats and attacks, the number of threat agents in the cyber world, the magnitude of their growth all over the world and increasing threats from various sources are ever changing.

Attack vectors and attack methodsmutate and grow in complexity as and when sophisticated and complex nature of ICT design components evolve.

'Landscapes of Cyber attacks' arethus identified by constantly updating information about context of incidences ofthreat, vulnerability and trust breach at various places.

Following aredomains where one can clearly perceive 'The Changing Landscape of Cyber Attacks'

• Banking and Financial Market.
• National Public Network.
• COTS(Commercial off the shelf) and GOTS(Government off the shelf) Hardware.
• Industrial Control and Power System infrastructure.
• Social Networks

Recently, the potential of machine and consequently deep learning for side-channel analysis was discovered and confirmed even on protected cryptographic implementations. The success of those experiments has led to deep learning techniques becoming a mainstream component in side-channel leakage evaluations. Orthogonally to this, some recent work has shown that neural networks can be reversed engineered by the side-channel attacker, i.e., the adversary using physical leakage such as timing and EM. This makes neural nets an interesting target as in some applications such as security evaluation, HD maps for autonomous vehicles and in all cases where optimized networks are considered an IP. This talk will survey this interplay of deep learning and side-channel analysis and its impact on embedded systems security.