CYBER-SECURITY AIDED BY

ARTIFICIAL INTELLIGENCE

ASC27 2022. REV 1.0 ENG

 2
What We Do
THE AI COMPANY “We Build Knowledge”

ARTIFICIAL INTELLIGENCE CYBER SECURITY

ASC27 2022 WWW.ASC27.COM

 3

30 PEOPLE TEAM
Offices in Rome, Milan, Bologna, and Chieti
Passionate people AVG HeadCount growth +10%/month
Multiple Award winnings

Our dream Entitled supplier of many large enterprises

Team
400 15 8
YEARS EXPERIENCE DEVELOPERS ARTIFICIAL INTELLIGENCE

ASC27 2022 WWW.ASC27.COM

 4
ARTIFICIAL ARTIFICIAL We develop ON them
INTELLIGENCE INTELLIGENCE

Silicon
ARTIFICIAL
INTELLIGENCE
ARTIFICIAL
INTELLIGENCE
PARTNERS

ARTIFICIAL ARTIFICIAL
INTELLIGENCE INTELLIGENCE

ARTIFICIAL
INTELLIGENCE

We develop WITH them

Software CYBER
SECURITY
ARTIFICIAL
INTELLIGENCE

PARTNERS

ARTIFICIAL
INTELLIGENCE

and
ASC27 2022 more... WWW.ASC27.COM
 5
Best Startup
Pitch WINNER

Awarded by
WORLD ARTIFICIAL
INTELLIGENCE
CONFERENCE 2021
WITH

WATCH VIDEO-

World TOP 10 BPAA

(Best Practice Applied Algorithms)
ASC27 2022 WWW.ASC27.COM
 Product Portfolio 6

ARTIFICIAL
INTELLIGENCE
CYBER
SECURITY

ASC27 2022 WWW.ASC27.COM

 7
HOMEBREWED SKILLS

We don’t RENT AI, AI VISION AI TEXT AI EMPATH CYBER

We DO AI for our
customers.
Safety Advice BoosterEye DocTracer
Surveillance Data 2 Text AICHO AuthSpine
Other Text 2 Text BlackWidow
Cognitive SecureTunnel

Customization and vertical developing capacity

from ground up. We can arrange our Cognitive
pipeline bricks to satisfy our customer’s needs.
AI Engineering and software skills in a single
Company. From Embedded systems to the
Cloud and beyond.

ASC27 2022 WWW.ASC27.COM

 8

SOLUTION
PROBLEM
Nowadays, all industry plants are plenty of Using AI BlackBox
unknown IoT and embedded devices.
Very often, the owners don't have the capability to Inspector, humans
inspect the source code of those devices.
BlackBox inspection is a time-consuming Activity.
can focus only on the
No one can check ALL the devices they had devices that need
attached to their network.
their attention.

ASC27 2022 WWW.ASC27.COM

 IOT DEVICES 9

TRL6

Security also in external devices: from device the firmware is extracted and
analyzed, identifying backdoors and security issues.
ASC27 2022 WWW.ASC27.COM
 ALL IN ONE 10

03. Language Lifting

- Convert sequences of 0 and 1 into a natural language
description and transform the code making it more
understandable and human-readable.

01. Firmware 02. Binary

Extraction Inspection
- Device based - Binary Analytics
- OTA Update based - Backdoor Database
- Repository Based - Vulnerability Database 04. AI Inspection
- Use AI to identify possible hidden threats in the
device internal functions. All in one platform.
ASC27 2022 WWW.ASC27.COM
 HOW IT WORKS? 11

IOT Binary Firmware High Level AI POWERED

Device Extraction Language Lifting search backdoor
The software analyzes and is The software extracts the Refinement of the binary code and subdivision into The AI analyzes the refined code and
able to read any device binary code of the firmware classes and categories according to the functions in autonomously detects backdoor or security
inside any circuit an High-Level language problems

ASC27 2022 WWW.ASC27.COM

 THE DETAILS 12

Information
Extrapolation
Extract and display information about the specified
firmware:

● Disassembled/decompiled individual function code

navigable and displayed in blocks.
● Various information from the binary (text,
constants, etc).
● Possibility to analyze the firmware directly from
our platform.

DEEP
IN DEPTH DE
SCRIPTION
OF THE FEATURE

ASC27 2022 WWW.ASC27.COM

 THE DETAILS 13

Analysis
Deep neural network-based behavioral analysis:

● Creation of ad-hoc models and heuristics for the discovery

of abnormal and undocumented behaviors in the specified
firmware.
● Creation of a functionality profile and detection of possible
deviation from the expected behavior.
● Alarming and analysis system for the detections mentioned
above.

DEEP
IN DEPTH DE
SCRIPTION
OF THE FEATURE

ASC27 2022 WWW.ASC27.COM

 THE DETAILS 14

Deep Neural
Network
Deep neural network analysis for backdoor detection with
automatic alarm:

● Creation of ad-hoc models for backdoor detection in the

specified firmware.
● Pre-existing backdoors.
● Firmware compromised in transit (update).
● Firmware compromised during use (modified by
exploitation of a device vulnerability).

DEEP
IN DEPTH DE
SCRIPTION
OF THE FEATURE

ASC27 2022 WWW.ASC27.COM

 Architecture Overview (Information Extraction) 15

TECHNICA
VERY TECHN
ICAL DESCRIP
TION OF THE
L
PROJECT

ASC27 2022 WWW.ASC27.COM

 Architecture Overview (AI Powered Analysis) 16

TECHNICA
VERY TECHN
ICAL DESCRIP
TION OF THE
L
PROJECT

ASC27 2022 WWW.ASC27.COM

 Architecture Overview (Information Extraction) 17

AI BlackBox
Traditional

Architecture Independent Analysis

Overall Data Analysed

Architectures Scalability

Features Scalability

Maintenance Effort

ASC27 2022 WWW.ASC27.COM

 ALL IN ONE 18

A
A
M
M A
A
FW A
M
A
FIRMWARE
M
MODULE
ANOMALY
A

ASC27 2022 WWW.ASC27.COM

 19

ASC27 2022 WWW.ASC27.COM

 20

THANK YOU

Q&A TIME

ASC27 2022 CEO Nicola Grandis nicola@asc27.com WWW.ASC27.COM