We love the new technologies here at MyFoodTrust, of course in relation to improving the current lack of transparency. Last week we talked to Daniel from Bext360, and their use of blockchain and AI. Today we focus on AI again, which we find super interesting as a tool for food transparency, so it was a no brainer to do a interview with Rufus from SensoMind.
Read here, how SensoMind have applied AI to create a system to detect anomalies in food products and what role AI will play in creating transparency in food supply chains in the future.
Can you start with telling us a little about yourself and SensoMind?Hi, my name is Rufus. I’ve been involved in AI ever since I studied at MIT in 2004 at their Artificial Intelligence Lab. I hold both a PhD and an MBA and have a passion for innovation management and entrepreneurship. Previously I worked a lot with perception for self-driving vehicles in agriculture. I founded Sensomind with my partner in 2016 in order to democratize AI and get it out to the masses. We’ve built our own set of tools around top AI products such as Googles Tensorflow which we thought at the time were too much targetting data scientists and not enough the engineers that are out in the field today. Our core competencies lie in analysis of complex sensor data. This is available in abundance in manufacturing so is one reason why we have gotten into this industry.
Sounds interesting, but can Sensomind’s AI technology be applied on food?
We’ve been working extensively with food manufacturing customers where our technology can be used for quality monitoring and sorting of food products. Most of our solutions are based on optical sensors (Such as cameras and multi-spectral imaging). Vision technology has been around in the food industry for 10-20 years but it’s been very difficult to apply it to food products with organic shapes and high variety. Examples include monitoring breads, meat, and fruits & vegetables. With AI you can teach the system just be showing it examples which opens up for completely new applications. An example can be automating the cutting of meat. The price of a final product has a large influence on the cutting being done correctly and it can be very difficult using traditional computer vision to recognize exactly where to cut.
And in relation to that, can Sensomind’s technology help tackle the problem of, e.g. food contamination or unapproved enhancements/additives in food?
So we have a system to detect anomalies which can for example detect contaminants. In the food industry we have used this to detect contaminants such as bone fragments, metal, plastic, and other objects which shouldn’t be there. Unlike a human operator, our system never tires. Unapproved additives is difficult to detect using traditional color cameras so here we work with spectrometers or multi-spectral cameras. Using traditional computer vision an engineer would normally sit and try to make a model for different additives based on a pre-conceived notion of what to look for. AI allows a more statistic and data-driven approach which reduces the chance of unapproved additives making it through the production undetected.
In your opinion, what role does new technologies, e.g. AI, play in creating transparency in food supply chains?
Supply chains are notoriously difficult to model because of large amounts of often poor quality or missing data. AI is really good at crunching numbers and extracting meaningful informations from poor quality and multi-source data (including images, text, numbers, etc). AI can help piece together the information about specific products which would be impossible to model by hand.
If someone was interested in learning more about the work you do, where can the find more?
The obvious thing would be to contact me. Check out our website (sensomind.com). We have a number of international projects going so location is often not a big issue.
A big thank you to Rufus, and great to hear of the use of AI in the food supply chain. Here at MyFoodTrust, we are always looking for how new technologies can enhance transparency and traceability.
So if you know of any startups, please let me know!
Welcome to Part 2 of my interview with Henrik Granau, or more correctly, Mr. RFID. If you haven’t read Part 1, I urge you to do that, to get a good understanding of Henrik, RFID and its challenges. Simply click here!
If you already have read Part 1, I love that you did, then you know that this part is going to focus on the opportunities with RFID within food traceability/transparency. No more small talk, let’s dive into it!
And then turning from the challenges, what are the general opportunities with RFID?
There are a lot of application areas and certain Industries where there are still obvious opportunities, but in general I believe it is in the combination with other technologies we see the huge potential;
By using RFID you create a transparency on how goods, assets etc. are flowing through your operation and if you add to this detailed information on how the workflow is actually being performed, you have established the foundation of making better decisions in your organization.
When combining these operational data with other data (‘Big Data’) and Software Robots (Artificial Intelligence), you can create new services and business models (‘disruption’).
Almost everything within the area of ‘Internet of Things’ involve wireless communication with a device which has to be uniquely identified to make sense – hence “RFID” will be ‘pervasive’.
Building on that, from your knowledge, what are the potential of RFID tags to create greater traceability of food products and why?
Internationally we already have a number of good cases in food traceability (Fish, livestock, vegetables) with RFID and combined with temperature sensors, we have established better cold-chain management. I believe that traceability within the supply chain can still be improved, but in general the technology is in place and we have the good cases with documented results.
The challenge is that we want the consumers to be able to have the complete history of each item available on their smartphone with one scan. If it’s RFID (NFC) or 2D barcode doesn’t matter so much – the challenge is to capture all the information during the product’s lifetime automatically which is achieved by using RFID on the transportation unit. For this to work, the unique item numbers which are packed has to be associated with the unique identifier of the transportation unit, and in some areas you will then have to add some evidence that the goods hasn’t been tampered with during transportation.
So creating greater traceability is possible with RFID, but you have other issues depending on the objectives; 1) for manufacturers to issue effective recalls, 2) for consumers to check the goods before consumption, 3) for protecting against fraud, etc.
What potential do you see in a RFID/Blockchain combination to create greater traceability of food products, from your knowledge?
I am not a blockchain specialist, but I understand that what blockchain can add is a bulletproof distributed verification mechanism. So, when the issue is to have verification that a specific organization is guaranteeing that their part of the traceability data are valid, then you could use blockchain to lock a certain ‘hand-over’ transaction with some associated data. If RFID is used then this process could be done automatically at choke points. As an anti-counterfeit method.
I believe I can learn more about the potential with RFID/Blockchain by being kept updated on your progress.
If someone was interested in RFID, what would be a few things you would suggest to investigate further?
I will start with recommending to vist www.rfididk.org. This is RFID I Danmark’s website where we have tried to give a good introduction to RFID – especially through cases and slides from presentations held at our conferences.
As a special service the RFID I Danmark Association is offering that anyone for free can contact me with initial questions. You can mail me at firstname.lastname@example.org or you can call me at +45 21 832 835.
Thank you to Henrik, for his great insights into RFID and the opportunities. From this and what I learned from the RFID in Denmark conference, I see great potential for a RFID/Blockchain solution in supply chains. RFID will secure correct data inputs, which can’t be tampered or adulterated, which then are data inputs for the immutable blocks in the blockchain application.
One of the key takeaways from the conference, was the lack of adoption and their one-sided focus on RFID being a inventory solution, and not grasping a more holistic picture of what the tech can do. And I feel that, that is a general thing when investigating new technologies, that it is very one-sided and not trying to connect all the dots.
As the second interview post I wanted to, hopefully, expand your knowledge about what tech can help with better traceability and transparency of food products. One of these tech’s is RFID. RFID is not new to scene of tech developments, but maybe looking to be a “revival”, due to its capabilities for easy tracking and tracing of products. So if you want to learn about RFID, from Mr. RFID himself, you have to read this.
I had the pleasure of meeting Henrik at the RFID in Denmark conference at IT University in Copenhagen this summer. I had touched upon RFID during the writing of my master thesis, and wanted to learn more about it’s potential in retail, and to know about where we are in terms of adoption by companies. To get those learnings, you have to read part 2. But for now, lets talk about RFID, here goes!
Can you start with telling us a little about yourself?
Well, I am what you would call an experienced executive, having conducted most of my career in international IT Companies, where I have developed my strategic outlook and knowledge for the successful Marketing, Selling and Implementation of high level and complex business solutions.
I was originally an IT Expert working with development of large complex IT Systems. As an example, I was Project Manager for the development of the world first distributed real time trading system for Financial Instruments (for Copenhagen Stock-Exchange 1986-1988). Since 1990 it has been General Management at C-level.
What are RFID technology and why is it important?
RFID is an abbreviation for Radio Frequency Identification – using radio waves to identify objects or people. This could be done on very long distances (Satelites, GPS etc.) and on very short distances (Access to buildings, wireless payments etc.). Some Frequency bands are in the regulation allocated to RFID use and a lot of people only perceive these as “RFID”, but in my definition it is everything using radio waves to identify – including WiFi, Bluetooth etc.
When RFID back in 2004-2005 was hyped as the next big thing – the replacement of barcodes – it was however very much focused on what we call “passive RFID”, where the term ‘passive’ means no power source on the RFID Tag itself (no battery). A passive RFID Tag is a very simple microchip with an antenna and only when it is in the proximity of an RFID Reader the chip is powered up (by the radio waves from the Reader) and it can do very limited operations, such as telling it’s unique Identifier.
Passive RFID Tags are now standardized and the prices has decreased to a level where it really make good sense to attach RFID Tags to single items. Because of the lack of battery, the lifetime of an RFID Tag can be considered like ‘forever’.
The technology is used across Industries and in a lot of different application areas. You can track a product throughout it’s entire lifetime establishing complete transparency in your business operation. Examples are Library books, Fashion clothes, containers, airplane parts etc. etc.
We have thousands of successful implementations – the technology is working and the cost is justified through achieved business benefits – but a lot of organizations have never learned about the technology in relation to their operation.
Great, but how did you first get involved with RFID?
I was introduced to RFID when I accepted the challenge to be heading a Danish start-up company, RFIDsec in 2005 – a company with a mission to set new standards for security and privacy in RFID. We developed our own security features at the chip level as well as at the solution level with end-to-end control. Unfortunately we had to close RFIDsec in 2010, but all our features are now a part of the new international standards which were finalized in 2015.
During my 5 years as CEO for RFIDsec (2005-2010), I established my international network within the RFID world – because I had to get involved in all formal as well as informal standardization activities around security and privacy issues with RFID technology.
I was a co-founder of the RACE Network (Racing Awareness and Competitiveness in Europe) who was advising the EU Commission in RFID matters 2008-2011. The RACE Network was later renamed to ‘RFID in Europe’.
What are your current work in relation to RFID?
In 2010 I founded the thematic network “RFID i Danmark” where I am still putting a lot of time and effort into nursing the initiative. Every year we are organizing the largest RFID Event in the Nordic countries – the 7´th of its kind was held in Copenhagen at the IT University on June 14’th 2017.
In the network I am known as “Mr. RFID” and since 2014 I have also taken on the challenge to build up a strong organization in the Nordic countries for AIM Global. I am also Vice Chairman of the Board at AIM Europe.
In addition to the networking activities, I work as an independent Management Consultant where most of my activities are in the area of tracking, tracing and locating.
I help companies select the most appropriate technologies and standards to their usage and keep myself updated on the technological development through a good relationship with the manufacturers and resellers of RFID products.
I am the guy who knows what is going on internationally and nationally within the area of RFID.
What is the most challenging for a general RFID adoption at the moment?
Well, in general Supply Chain Management, in Fashion clothes and apparel, in Libraries, in ticketing, access cards etc. I believe that we on an international level actually has reached ‘general RFID adoption’. If you look at the hype created back in 2004-2005, where RFID was predicted to be a general replacement of barcodes, though I will still state that this will never happen with the current silicon based RFID technology. It just doesn’t make sense to put RFID tags on each item of bubble gum, milk etc. – the total cost of attaching an RFID Label is still 30-40 times the cost of using a traditional barcode.
Across Industries I would however still claim that lack of knowledge is the most important reason for not implementing RFID.
Want to read about the opportunities that Henrik sees with RFID in supply chains? RFID’s possibility to create better traceability in food products? And Henrik’s thoughts on a RFID/Blockchain combination to greater traceability in food products?
The blockchain technology was invented by a person under the alias Satoshi Nakamoto, to support the cryptocurrency Bitcoin (Nakamoto 2007). For the first time it was possible for many users to trade values with each other over the Internet without the need for a third party or intermediary – typically a bank – to verify the transaction. A blockchain is a ledger of facts, replicated across several computers assembled in a distributed peer-to-peer network. Or put simply, a chain of blocks (Beck 2017). Anyone participating in a blockchain can review the entries in it; users can update the blockchain only by consensus of a majority of participants. Once entered into a blockchain, information can never be erased (Nakamoto 2007: 2).
Blocks are an order of facts in a network of non-trusted peers, similar to how Uber’s technology intermediates between suppliers and consumers of transportation. Facts are grouped in blocks, and there is only a single chain of blocks, which then is replicated in the entire network. Each block has a reference to the previous block, through the hashing cryptography that links the blocks. Some of the nodes in the chain create a new block with pending facts. They, in the case of bitcoin miners, compete to see if their local block is going to become the next block in the chain for the entire network, called proof of work. Then this block is sent to all other nodes in the network. All nodes run a check on that to see if the block is correct, then add it to their copy of the chain, and try to build a new block with new pending facts (Nakamoto 2007: 3).
But it has gradually become clear that the technique has much broader applications than just acting as the backbone of Bitcoin. One of the key elements is the ledger, which is a database of the content of the blockchain – whether it is bitcoin transactions, intelligent smart contracts, or something else (Boye 2016).
Blockchain is a type of electronic ledger created to ensure that once a party transfers a digital asset, he cannot transfer it to anyone else, prevent double spending. Unlike other ledgers, blockchain is owned by its participants, and decisions about what it records are subject to participant consensus.
Recording accuracy is ensured by duplication: every participant has a copy of the ledger. Discrepancy-resolution mechanisms ensure that all copies reflect an identical history. Though permissions can be managed with a fair degree of control, by default any permitted participant can view all transactions. Thus together with immutability, notarization and assured provenance, transparency is a core blockchain attribute (1).
There are many ways of applying a blockchain technology, in short, either as a public blockchain, a private blockchain, or as a consortium blockchain. A public blockchain is a blockchain that anyone in the world can read, through which anyone in the world can send transactions, and include transactions if they are valid, i.e. Bitcoin (Buterin 2015). A fully private blockchain is a blockchain where write permissions are kept “centralized” to one or few institutions, i.e. banks (Buterin 2015). A consortium blockchain is a blockchain where the consensus process is controlled by a pre-selected set of nodes. An example, is a consortium of 15 financial institutions, each of which operates a node and of which 10 must sign every block in order for the block to be valid. A consortium blockchain can be altered to fit the need of the one using it, ex that the R3 consortium want different “rules”, than the Hyperledger consortium or Ethereum Alliance (Buterin 2015; R3; Hyperledger).
Public blockchains can offer advantages that a private blockchain and consortium simply cannot, and vice versa. The take-away with the different ways of adopting blockchain technology, in relation to COOP Trading, is what they want to gain from a blockchain solution, who should be a part of it, who should have read and write permissions and what data can’t be shared. One must have a high due diligence in order to research the possibilities and challenges with a blockchain solution.
“A block is the ‘current’ part of a blockchain which records some or all of the recent transactions, and once completed goes into the blockchain as permanent database. Each time a block gets completed, a new block is generated. There is a countless number of such blocks in the blockchain. The blocks are linked to each other (like a chain) in proper linear, chronological order with every block containing a hash of the previous block.” (Investopedia)
Finally, blockchain isn’t simply a secure collective database. In addition to transactions, it also records and executes simple programs.
The idea of pre-programed conditions, interfaced between users, and then broadcasted to everyone, is called a smart contract. A contract is a promise that signing parties agree to make legally-enforceable. Proponents of smart contracts claim that many kinds of contractual clauses can be partially or fully self-executing, even self-enforcing, or both. The aim of smart contracts is to provide security, which is superior to traditional contract law and to reduce other transaction costs associated with contracting (Tapscott 2016: 105-108). Buterin explains it as: “(…) then we can cut costs to near-zero with a smart contract.” (Parker 2016).
Blockchain smart contracts may also influence, or be influenced by, product movements. For instance, a positive QA test indication can release a part for assembly. However, today that role is played by ERP systems. Blockchain technology doesn’t necessarily add value in such traditional operations management tasks (1).
“An asset or currency is transferred into a program and the program runs this code and at some point it automatically validates a condition and it automatically determines whether the asset should go to one person or back to the other person, or whether it should be immediately refunded to the person who sent it or some combination thereof.” (BlockGeeks)