Artificial Intelligence – Claudio R&D +++ DIstributed Ledger
Claudio & co. has doing Research & Development in Artificial Intelligence and Machine Learning for many years. Focussing on Deep Learning with aspects from Neuroscience for analysing BigData towards AGI, as well as Natural Language Processing for improved and people-centric Human-Computer-Interaction. We like to solve the connection between World-knowledge databases (symbolic) and statistical machine learning.
The current quest is to gather data, which is more and more coming from IoT not only on a Blockchain, but encrypted on a next generation Distributed Ledger.
www.ai-claudio.com is engaged in research, authoring articles, blogging and connecting the communtiy.
Please feel free to browse through the posts or to get in touch through the contact form below.
MHP – A Porsche Company
MHP is a leading consulting firm in digitalisation and automotive and a subsidary of Porsche.
Nonetheless MHP is consulting many other companies then Porsche AG as well, like 6 automotive manufacturers, 90% of the German Top 25 automotive suppliers, a third of all German Top 100 commercial enterprises (by turnover) …
For inquiries about Consulting for Artificial Intelligence, Big Data Analytics, Connected Car or Industry 4.0 with Blockchain – feel free to contact Claudio @ Big Data Analytics in MHP or via phone: +49 151 4066 7937
The digital world is moving towards distribution faster then we are realising it. The World Economic Forum published in a study 2016 even that in 2027 at least 10% of the global GDP is processed with the Blockchain technology.
Interestingly the distribution of data, which is based on consensus of the participants, does not divide data – is actually brings merge of various data sources even closer. It’s like in today’s world our friends and family become far more distributed in all places all over the world – yet the world seems to become smaller – as we connect to friends from all over.
The future of Blockchain is not Blockchain.
Of course Blockchain mostly introduced us to a series of new technologies – which are not merely ‚Chains of blocks‘ but rather utilise the corner stone ideas of Blockchain and Bitcoin.
What are those is dependant of each Use Case, including the data storage across all places, transparency, immutability, trust in trust-less, democratisation, incentivation, durability ….
Therefore the more general term Distributed Ledger Technologies (DLT) might be more accurate.
GDPR & Data Privacy loves Distributed Ledger Technologies
The initial paradox of data for the eternity and the data privacy including the right to be forgotten for personal data is actually a core strength of this technology. While data can be stored encrypted and decentralised, the access and permission log can be implemented immutable – and therefor give the user and data owner full transparency over their data.
IoT – Blockchain – Artificial Intelligence
As more and more sensor data is generated and stored in a public database and accessible by more and more users, the benefit of A.I. increases with this data amounts for training.
As a Feature Speaker I will be at the Machine Intelligence Exchange in Berlin on June 20th 2017. I will talk about Deep Learning for Internet-of-Things such as Smart Cars.
Please inform me if you are coming as well or want to have a talk on the day.
At a later stage I will share some of our presentation here too.
While in Europe and US the classic car manufactures created often one-to-one partnerships with OEMs and tech companies in the past, in Asia now also new business models are rising:
Some previous partnerships include (in alphabetically ordered):
- Audi Autonomous Car in 2020 with NVIDIA
- BMW partners close with Delphi Automotive and Intel, which now also owns Mobileye
- Bosch partners with NVIDIA and Daimler as well as several car manufactures
- Ford bought SAIPS Machine Learning, co-invested in Velodyne (with Baidu) for it’s 2021 robot cabs agenda
- Here partners with Mobileye and NVIDIA
- Hyundai Motor presented the Ioniq Concept
- Japan created an autonomous car consortium for the 2020 Olympics in Tokyo
- Samsung Electronics for Automotive bought tech with Harman
- Volvo struggled with Uber (link)
- Volkswagen online Mobileye and Delphi in 2018
- ZF and NVIDIA collaborate on embedded autonomous driving
Currently manufactures and OEMs see their need of software, hardware and data. Therefore building partnerships conventional one-to-one or one-to-few. (Additionally past process optimisation often came from automotive – now also the IT offers more and more team process optimisation, such as with SCRUM and Agile.)
But besides the Google car, which was in the news for years, Baidu might be even closer to Googles regular market approach:
Open software and services and creating a ecosystem.
„Baidu, which is China’s Google, is leading the effort with big data and deep learning and recently began public trials of its own self-driving cars, despite severing its autonomous driving partnership with BMW, noting the two companies differed on how to proceed.“
Baidu however, goes a step or two further, trying to build an open source like autonomous car reference implementation – like Google ownes Android.
Their advantage is, like Google, they don’t have a car business to loose, yet to gain. This can be a whole market disruption, as many new players on the market can arise utilising their technology.
Hardware: Eventually besides the software algorithms maybe hardware will be manufatured within the mainland Chinese ecosystem as well.
For example the top #1 fastest supercomputer in the world, Sunway TaihuLight in Wuxi runs already on only local processors with with 93 petaflops.
Read about Baidu: http://usa.baidu.com/adu/
„Ford’s research has found that 84 percent of Indians and 78 percent of Chinese could see themselves owning an autonomous car, compared with only 40 percent of Americans and 30 percent of Britons, Connelly said.“
Hence, form my point of view: Tech soft- and hardware suppliers and car manufactures / OEMs have to collaborate more, invest heavily in Deep Learning and see openness as an advantage.
Additionally, have a look at a Ptolemus Partnership Map:
For all my international colleagues, let’s meet in Berlin on June 1st or the days before.
After the keynote panel talk and the break I will give a speech about these topics:
Give me a call or e-mail in case you are in town!
As responsible for the topic and team of Data Scientists in our business area, we moved to using Deep Learning more and more instead of the common classic statistical algorithms.
Of course there are a tons of different Frameworks but I ♥ keras
I came to I love keras as the meta language is able to work on top of TensorFlow, theano and also Microsoft CTNK is in development of supporting it.
Even for some, and I believe only a few, use cases the pure frameworks are more powerful – in most cases development time, optimisation, collaboration and education with it is amazingly efficient.
Read about keras: https://keras.io
Additionally, also Lasagne is such a good meta language approach. As I didn’t implement it yet in any projects, I can’t compare it much.
I will give a deeper insight in my following speeches and webinar.
In the last time a lot of great AI core software came out.
TensorFlow by Google integrated the code of DeepMind and is strong in imae recognition as well as natural language processing.
Besides the OpenAI group is working on a open and bigplayer independent AI platform for the masses. Of cource also Microsoft is strong with Machine Learning on Azure and Amazons DSSTNE is still living a shaddow life.
Within Automotive there are news everyday about new accomplishmens. Like from Audi, Mercedes, Telsa and all the others. Also IT gigants like Google and open-secretly Apple as well as often critizised Uber.
Even there are many restrictions often commented – mostly within the areas of ethics and laws – it’s a clear trend towards autonomous driving with more and more driver assistance systems.
Already for me, as digital native, the software features running within a car trumps the design & interrieur etc.
Working within the Automotive area on machine learning and artificial intelligence means having great changes ahead.
So more will follow 🙂
If you see a button with the description ‘auto’ on a machine – you know you don’t have to do much – the machine will work it out itself – and that will soon be the same for cars.
In Germany we already call a car them ‘Auto’ which is a short for ‘Automobil’, deriving from the greek αὐτός and latin ‘mobilis’ – “self moving”. Well the name has it already in it – and this year there were so many news about the selfdriving cars.
Daimlers concept car F015 has some nice elements for people inside and outside the car.
Will the be no traffic jams if all cars drive in a steady way, with a secure but short distance to each other without any surprising behaviour? No traffic lights needed when cars are connected? Order a car with a desk during daytime or a comfy couch on the evening. No need of an all purpose car – enjoy a nice sporty one if you feel like it – or an super energy efficient if you prefer.
As soon as most traffic is done by autonomous cars – or as I like to call them Auto Autos safety will increase incredible – and small cars will be as nearly as safe as huge ones.
Talking about machine learning today doesn’t go without talking about distributed fault-tolerant data storage and query system.
Hadoop is an open-source Apache Project derived 2006 from Yahoo! and based on papers from Google in 2003, is such a widely used system. Even though it is basically ‘mostly just’ a filesystem it’s three biggest advantages are
At the core of Hadoop is the filesystem HDFS (Hadoop Distributed File System) which stores it’s data in blocks across all DataNode machines. The data is replicated usually on a (or more) machine in the same rack, as well as on an other rack. Clients connecting to Hadoop to read or write will first question the NameNode which will tell them at which DataNodes they can attempt to access.
The current Hadoop 2.x version rely on YARN (Yet another Resource Negotiater) for that and with data and server replication there is no single-point-of-failure anymore. On top of that Hadoop uses MapReduce as key/value database. Therefore Hadoop is great for lots of data retrievals and querying.
It’s drawbacks are: the data in HDFS is not editable, only append able, it takes a lot of configuration work and without any additions and it’s not for real-time queries.
Artificial Intelligence is a huge topic at the ongoing Google I/O 2015.
Life from the keynote: most products discussed have new features heavily dependant on A.I.!
All-New Google Photos is using improved face and pattern recognition. Not only the face recognition was improved, which was already available in Google Picasa and other apps. But additionally and new to consumer photo album software – it also will use pattern recognition to add automatically searchable text tags.
Additional automatically cut and edited videos known as “auto awesome” will be improved as Google Photos “Assistant” – using A.I. internally to figure out which video sequences and photos to use.
Google Now has more then 1 billion information pieces able to show users to assist them at their current action. Simple via (A.I. powered) speech-recognition connected information are shown.
They have mentioned that since 2013, when speech recognition had a failure of about 23%, it’s down today to 8%. So we can expect this number to go down even further.
Len Epp is writing in A Vision Of A Driverless Future | TechCrunch about conclusions of the possibility of autonomous cars. Not only the drivers habits will change but the whole industry: shop on wheels, local services, no ownership of cars needed anymore, huge variation in sizes, …
So the AI powered self-driving cars will bring a huge change in our everyday life and for many businesses – great potential for many startups too.
Instagram Feed: #mhpclaudio[instagram-feed num=9 cols=3]
Image Sources: All rights are at Claudio Weck, or CC0 contributors, as well as pixabay.com, except those from Shutterstock.com, like:
Christian Lagerek https://www.shutterstock.com/de/image-illustration/human-brain-communication-computers-motherboard-605481167