2. Challenge 2
Free up personnel through the application of
innovative use of machine learning algorithms
and artificial intelligence (AI) for military
advantage
3. Next generation Air Force
Information
collection
Human
analytic
capacity
People
TechnologyProcess
Challenge
4. Decision advantage
Manage, analyse and exploit
multiple information sources
…….at pace
Exponential data
Identify the right 1%
Constrained
human capacity
6. Exponential data – ISR Services
Multi-Intelligence
Fusion & Cross-Cue
Automation
AI
Analytics
Optimise
Intelligence
Analyst
Fusion /
Cross-Cue
Imagery
Multi-Spectral
Hyper-Spectral
Electronic
Communications
Foreign Intel Systems
Measurement & Signatures
Cyber & EM
Acoustics
Human
Open Source
Historical /Archive
Direct
Collect
Process
Disseminate
PROCESS Information = Human /
Machine Partnership
Decision
Advantage
8. Wider opportunities
Engineering and logistics
• improve aviation safety
• keep aircraft in the air for longer
• environmental stress and trend analysis
• work closer to mandated tolerance limits
Cyber defence
• continuous activity on networks
• identify the anomalies
9. Conclusion
• decision advantage
• exponential data vs human
capacity – close the gap
• the right 1% ..... at pace
• human and machine in
partnership
11. OFFICIAL
UK OFFICIAL
LSVRC* classification challenge:
error rates by year
red line = human error rate
…
Face recognitionSpeech recognition Lip reading Machine translation
*Large Scale Visual Recognition Challenge
12. OFFICIAL
What do we want?
Over-fitting
Free and open-
source software
(where appropriate)
Solve one aspect of
the problem well
13. OFFICIAL
Automated activity classification
MOD requires methods for automated detection and classification of
activities and intents from multiple sensor types using state-of-the-art
machine learning and artificial intelligence (AI)
Fathom neural computer stick
Adversarial machine learning example
• beyond simple feature extraction
• ability to operate “at the edge”
• semi-supervised and un-
supervised methods
• approaches to enable robust
deployment (for example
adversarial machine learning)
14. OFFICIAL
Cognitive computing
UK OFFICIAL
Automated speech recognition
Knowledge graphs
Natural language question answering
Automation of manual
tasks
Flag adversary activity of
interest
Infer new “knowledge”
Identification of false
information
15. OFFICIAL
Combined human/machine derived models
UK OFFICIAL
MOD is interested in the combination of human-
derived models, exploiting domain knowledge using
a rules-based approach; with machine-derived
models, which require large volumes of data and
driven by machine learning technologies. How do
we:
• combine data and human derived models
• build more robust statistical models of subjective measures
(for example assessment of threat)
• ensure data-driven models are transparent and
understandable for analysts and operators?
16. OFFICIAL
Predictive analytics
Application of machine learning in support of predictive modelling to guide military decision
making. MOD requires solutions which go beyond enhancing military understanding of
current situations, but predicts future outcomes, including actions, anomalies, intent and
movements, to guide decision makers in support of operational planning.
UK OFFICIAL
Information
overload
Situation
understanding
Predictive
analytics
Prescriptive
analytics
19. Challenge 3
To make effective use of operator cognitive capacity, particularly by human-
machine teaming
Key points for the Land Environment
• considerable improvements need to be made in the interaction between people
and systems
• develop approaches that enable collaborative decision making and
intelligence analysis to support planning activities and military operations
20. Real world considerations
• we start from a brownfield site
• need to straddle multiple branches
• data is everywhere but what matters most?
• there is no intelligence but information of specific value
• essential enabling conditions & foundations?
• we are still talking about the chaos of war
• our enemies have a very real vote
• our ability to operate over degraded networks and
federated command and control
21. Army considerations
Resetting focus to warfighting at Divisional
level:
• bandwidth, computation and size, weight
and power (SWAP)
• Moore’s Law and narrowing of technical
competitive edge
• international by design
• being a people AND platform force
• maximizing people and talent:
knowledge, skills and experience
22. Mission threads
• look beyond information exchange
requirements (IERs)
• gaps in our staff process/approach
• information must be treated and
consumed as an essential service
• must be command-driven and
anticipatory
23. Human information interaction
OFFICIAL
How can I (and my team):
• rapidly and intuitively locate key information for my role
• indicate that certain information is important, and why and when so I
can find it again
• record/create information without worrying where it is located and
not being able to find it again
• record key relationships between information
• understand accuracy and provenance
• be told if I need to know but don’t have permission to access
• prevent being swamped by the scale and complexity of available
information
24. What is the enabling
architecture in the fixed
space and deployed?
25. Wider Defence Lines Of Development
(DLOD) considerations
• personnel – what key skills and experience do we develop?
• doctrine – can we conceptually keep pace?
• infrastructure - what is the technology readiness level (TRL)
‘aiming point’?
• training
• individual, professional and collective burden?
• TRAIN AS WE FIGHT
• integration – let’s not be afraid to fail
• interoperability – designed in at the outset
27. OFFICIAL
Aims
OFFICIAL
Obtain and exploit innovative ideas that:
• Ensure that human cognitive capacity (which is limited) is applied to
those parts of military problems that humans can undertake best
• Reduce unnecessary consumption of human cognitive capacity on
activities better supported by automation
• Achieve the above by ensuring that human and automated parts
work effectively in unison avoiding pitfalls and problems
31. OFFICIALOFFICIAL
Summary
We are interested in solutions:
• which take account of team context
• that don’t increase training load, are intuitive to use, and adoptable by
non-experts operating in stressful environments
• that can start small and simple, have rapid application, but have the
potential to scale up
We are not interested in solutions:
• that replace the human component or relegate role of the human
• which fail to take account of identified automation pitfalls
• which might force people into unnatural ways of operating
• that are stand-alone human machine interaction technologies
33. OFFICIALOFFICIAL
Reasoning
Record and process reasoning related information
• represent/store questions, hypotheses, assumptions and
uncertainties
• continuously check reasoning against incoming data stream
• apply reasoning to generate new findings, create new
questions and hypotheses etc.
34. OFFICIALOFFICIAL
Relevant roles
Illustration by Andrew Rae
Tendency to automate everything or roles which humans can
do better
• for example abstraction, pattern matching across diverse input, self
assessment/reflection, idiosyncrasy, creativeness
Interested in
• novel approaches which demonstrate more appropriate assignment
of relevant tasks/roles to human and machine
• approaches which keep human interested, engaged and workload at
appropriate level (no under/overload)
Overall Concept
• team design based on SQEP of human and machine parts
35. OFFICIALOFFICIAL
Individual and team interaction
Tendency to stove-pipe human machine tasks/roles
• no effective team-working between human and machine
• teaming ‘capacity/behaviours’ is difficult
Interested in solutions that
• improve interworking based on a equivalent team member interaction
concept
• exploit team contextual information
• dynamically vary what human/machine parts are doing
Overall concept
• augment human teams with machine team members
40. OFFICIAL
Fast track
Phase 1
Project duration 3 months
Proposal up to £150,000
Phase 2
Project duration 6 months
Competition structure
Standard track
Phase 1
Project duration 6 months
Proposal up to £100,000
Phase 2
Project duration12 months
Same level of phase 2 funding between the two tracks
41. OFFICIAL
Fast track
• Higher level of phase 1 funding
• Shorter time to market
• Potential access to demonstration opportunities
Benefits
49. OFFICIAL
What we are interested in:
• integration of raw data sensors
• integration of intelligent information sources
• processing
• fusion
• autonomous sensor management
Challenge
50. OFFICIAL
What we are not interested in:
• mechanisms to enable non-cooperative access
to collection assets
• solutions where the number of sensors is limited
• distributed architectures
Challenge
51. OFFICIAL
Free up personnel by the
innovative use of machine learning
algorithms and artificial intelligence
for military advantage
Challenge
52. OFFICIAL
What we are interested in:
• automated activity classification
• cognitive computing
• combined human machine derived models
• predictive analytics
Challenge
53. OFFICIAL
What we are not interested in:
• machine-learning solutions which are highly
optimised for input training data, leading to
problems associated with over-fitting and
failure when environmental parameters
change
Challenge
55. OFFICIAL
What we are interested in:
• memory
• reasoning
• teaming – relevant roles
• teaming – individual and team interaction
Challenge
56. OFFICIAL
What we are not interested in solutions that:
• replace the human or which require no
human involvement
• are overly complex, require substantial
training
• force people into unnatural ways of
operating or behaving
Challenge
57. OFFICIAL
What we are not interested in solutions that:
• don’t include integration with other proposed
solutions delivering information and
processing capability
• use static information visualisation solutions
Challenge
68. Context
Operating impartially in the pre-competitive space allows an
excellent opportunity for open customer engagement and for UK
industry to collaborate and innovate effectively.
70. A proven partnership
• DGP Innovation Challenges
• Training
• Persistent Surveillance
• Big Data and Autonomy
• Designed to address
exportability and
exploitation
• £10M initial investment by
MoD
Enabling Exploitation
Co-Investment Delivered
MoD ATI Industry
71. Protection
Power
Communications
Data
Lower Cost of Ownership
Human Performance
Mobility
Lethality
Situational Awareness
Energy & Energy Distribution
Autonomy
Big Data
Communications
Low Cost Space
Materials & Manufacturing Technology
Military Aircraft
Quantum
Security
Sensing
Services
Training & Simulation
Systematic exploitation
75. DUTE Partnership with non-defence industry
DUTE is the DGP’s £20M Dual Use Technology cluster:
• DUTE was created to identify and leverage technologies from adjacent sectors such as rail and civil aerospace, and put them to
dual use. The initial cluster was founded with £13m of joint Government and Industry AMSCI funding
• through SME, Prime and mid-tier engagement, DUTE has raised further investment with adjacent sector co-funding of £7.5m, to
stimulate productivity, prosperity and export agendas in line with BEIS, MOD and DSO policy
Communities of Interest
SME Equity Fund
Dual Use Technology…
Developing UK Industry
Developing skills
with SME’s & large
companies
together; building
enduring value
chains
Creating the right
conditions to invest.
Embracing the
Defence Innovation
Initiative
Leveraging from non-
defence sectors.
Open engagements
through
Communities of
Interest
Exportability Training
Systems Engineering Masters
Apprenticeship Programme…
Innovation Challenges
Co-Investment Framework
Winning Exports
Understanding our
strategic markets &
approaching them
in a joined up
manner
Creating the most
capable Industry-
Government Teams
Strategic Market Analysis,
Country Engagement Plans…
Team UK
76. How can DGP support Defence, Security to unlock adjacent sector opportunities in
Defence Innovation and Industry growth?
Energy and Energy Distribution
Military Aircraft
Big Data
Communications
Low Cost Space
Materials and Manufacturing Tech
Autonomy
Quantum
Security
Sensing
Services
Training and Simulation
Protection
Power
Communications
Data
Lower Cost of Ownership
Human Performance
Mobility
Lethality
Situational Awareness
We can offer support and collaboration through the DGP communities of interest via the UK Defence
Solution Centre and DUTE in order to explore how relationships, independent from the contract with the
MOD, can maximise opportunities for Defence exports or sales into adjacent sectors.
77. A worked example of Dual Use Technology Exploitation currently under review
2015 20172016
Suppliers
Established
Team Build &
Support
Dual Use
Success
Dual Use
Export
Growth
August – November 2016
Adjacent Sector Exploitation
and Application
August 2015
DUTE Consortium
Building for UK Supply
Chain
May – July 2016
Engaged Support for
Submission and Review
DSC Persistent
Surveillance Challenge
Launched
September – October 2015
Opportunity Mapping with Zephyr Team
September 2015
DUTE Funds Launched @
DSEi
DSC Persistent Surveillance
Challenge Winners
Announced
DUTE Sector
Support
February - April
2016
Aligning non-
Defence R&D
January – November 2016
Aerospace & Automoive Partner and Engagement during project development
Automotive Capability Aligned to Support
Defence Markets
78. Potential dual use opportunities for the AI & Machine Learning which ca n leverage
commercial technologies inward to Defence . .
• THE COMMERCAIL NETWORK OPPORTUNITY
Britain is 54th in the world for 4G coverage with black
spots occurring in places that should have adequate
signals such as rail routes, roads and city centers . ... 5G
is coming & UK lead the innovation
• THE ENDLESS DEMAND FOR CONNECTIVITY
Connectivity in personal devices enabling greater safety,
security and maintenance scenarios
A sensor that communicates with other connected
service providers and devices to deliver relevant and
convenient digital services
• A COMMON OPPORTUNITY PRESENTED
Build a value chain grounded in innovation giving fast,
reliable, secure bandwidth so to unlock a UK
competitive advantage for defence and security
through commercial sector reuse
79. DGP support to the Defence Challenge FOR Innovation
• The Defence Growth Partnership, as part of the Government’s Defence Industrial Strategy: By
working with UK-DSC and DUTE we can offer independent support to:
• bring together existing capability from extensive interest groups for UK Defence
• drawing together the conversations and help foster focussed support from across sectors
• maximise focus and galvanise the engagement with DSA for growth and export
Communities of Interest
SME Equity Fund
Dual Use Technology…
Developing UK Industry
Developing skills
with SME’s & large
companies
together; building
enduring value
chains
Creating the right
conditions to invest.
Embracing the
Defence Innovation
Initiative
Exportability Training
Systems Engineering Masters
Apprenticeship Programme…
Innovation Challenges
Co-Investment Framework
Winning Exports
Understanding our
strategic markets &
approaching them
in a joined up
manner
Creating the most
capable Industry-
Government Teams
Strategic Market Analysis,
Country Engagement Plans…
Team UK
DGP support to the Defence Challenge FOR Innovation
Leveraging from non-
defence sectors.
Open engagements
through
Communities of
Interest