{"id":1201,"date":"2021-02-13T09:49:43","date_gmt":"2021-02-13T09:49:43","guid":{"rendered":"http:\/\/janbosch.com\/blog\/?p=1201"},"modified":"2021-02-13T09:49:50","modified_gmt":"2021-02-13T09:49:50","slug":"bringing-ai-to-the-edge","status":"publish","type":"post","link":"https:\/\/janbosch.com\/blog\/index.php\/2021\/02\/13\/bringing-ai-to-the-edge\/","title":{"rendered":"Bringing AI to the edge"},"content":{"rendered":"\n
\"\"
Image by Pete Linforth from Pixabay<\/figcaption><\/figure>\n\n\n\n

It seems that much of my work these days is concerned with bringing AI to the embedded-systems domain, understanding what the implications are and how companies should deal with it. In the discussions with technical experts and business leaders, however, I constantly run into several misconceptions.<\/p>\n\n\n\n

First, there are still several people out there who think about ML\/DL\n models as single-point solutions for a specific problem where we can \nbuild the model, integrate it, deploy the resulting software and be \ndone. Although one could operate in this way, it\u2019s basically ignoring \nthe vast majority of benefits that you could deliver. Instead, it\u2019s \nabout continuous everything. It\u2019s about a constant flow of software \nupdates to the field, a constant flow of data coming back from your \nsystems, constant retraining and updating of models. We all read about \nDevOps, DataOps and AIOps, but that really is what it\u2019s all about. The \nshorter you can make the cycles, the more value you can deliver.<\/p>\n\n\n\n

Second, many view AI as a technical challenge to be solved by R&D\n but unrelated to the rest of the company. The fact is that it will \nchange your business model. If you\u2019re currently using a transactional \nbusiness model where you get a one-time payment when you sell the \nproduct but need to provide software updates throughout the economic \nlife of the product, you\u2019re in trouble as your cost is continuous while \nyour revenue is not. The only way to get around this is to align your \nrevenue model with your cost model. For most companies that I work with,\n it means combining the upfront product sales with a continuous service \nrevenue model. This is actually a great way to generate more revenue \nfrom your customers.<\/p>\n\n\n\n

Third, especially in companies that build large, complex and \nexpensive systems, such as in automotive, telecommunications or \nautomation, there\u2019s no clear definition or understanding of the actual \nvalue provided to customers. The whole package sells and consequently, \nthe whole package is valuable. When moving towards continuous \nimprovement of systems, you need to choose where to focus your energy \nand time. And this requires you to understand what\u2019s actually valuable \nin your offering as improving commodity functionality using ML\/DL models\n or otherwise doesn\u2019t help your customer. We\u2019ve worked quite a bit with \ncompanies on modeling the value of their solutions or products to \ncustomers and it\u2019s surprisingly hard to be precise and concrete.<\/p>\n\n\n\n

Fourth, you can\u2019t use AI in isolation. Machine and deep-learning \nsolutions require data for training and operations. The data is \ngenerated by software that instruments the system. And of course, ML\/DL \nmodels are themselves software. The only difference is that AI software \nis programmed by data and pattern recognition, rather than by humans in \nan algorithmic fashion. This means that you need to be good at all \ndigital technologies.<\/p>\n\n\n\n

The final misconception is that many believe that once they have \ntrained an ML\/DL model and it performs well in prototyping, the hard \npart is done. This couldn\u2019t be further from the truth. The easiest part \nis actually creating a model and the hard part is industrializing it. \nIndustrializing means setting up the data pipelines, putting monitoring \nand logging in place, ensuring correct (or at least acceptable) behavior\n in all cases, developing solutions for DevOps, DataOps and AIOps, and \nso on. In an earlier post<\/a>, I wrote about our AI engineering research agenda where we capture the major challenges to be addressed.<\/p>\n\n\n\n

AI is moving to the edge \u2013 and it should because it has the potential of delivering enormous value. The challenge is that there still are quite a few misconceptions out there about what this actually means in practice. I\u2019ve discussed five of these and provided my viewpoint. AI belongs on the edge, but there\u2019s a lot of work around it that needs to be put in place at the same time. And as it won\u2019t happen by itself, it\u2019s critical that we get going on this yesterday.<\/p>\n\n\n\n

To get more insights earlier, sign up for my newsletter at <\/em>jan@janbosch.com<\/em><\/a> or follow me on<\/em> janbosch.com\/blog<\/em><\/a>, LinkedIn (<\/em>linkedin.com\/in\/janbosch<\/em><\/a>), Medium<\/a> or Twitter (<\/em>@JanBosch<\/em><\/a>).<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"

It seems that much of my work these days is concerned with bringing AI to the embedded-systems domain, understanding what the implications are and how companies should deal with it. In the discussions with technical experts and business leaders, however, I constantly run into several misconceptions. First, there are still several people out there who … Read more<\/a><\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"generate_page_header":"","footnotes":""},"categories":[15,10],"tags":[],"_links":{"self":[{"href":"https:\/\/janbosch.com\/blog\/index.php\/wp-json\/wp\/v2\/posts\/1201"}],"collection":[{"href":"https:\/\/janbosch.com\/blog\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/janbosch.com\/blog\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/janbosch.com\/blog\/index.php\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/janbosch.com\/blog\/index.php\/wp-json\/wp\/v2\/comments?post=1201"}],"version-history":[{"count":1,"href":"https:\/\/janbosch.com\/blog\/index.php\/wp-json\/wp\/v2\/posts\/1201\/revisions"}],"predecessor-version":[{"id":1203,"href":"https:\/\/janbosch.com\/blog\/index.php\/wp-json\/wp\/v2\/posts\/1201\/revisions\/1203"}],"wp:attachment":[{"href":"https:\/\/janbosch.com\/blog\/index.php\/wp-json\/wp\/v2\/media?parent=1201"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/janbosch.com\/blog\/index.php\/wp-json\/wp\/v2\/categories?post=1201"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/janbosch.com\/blog\/index.php\/wp-json\/wp\/v2\/tags?post=1201"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}