A curated publication on applied Machine Learning, Data Science & AI. Enjoy technical long-reads, case studies, interviews and coding tutorials.

AIgents acquires Elify.io

We are proud to announce that AIgents has acquired Elify.io.

Read the full story

Hackathon on improving seedling classification performance through advanced feature engineering

In this first Hackathon of 2020, a group of Itilians, data science students, and other data enthusiasts came together to sharpen their data science skills. The topic: improve the accuracy of the current classification model used in a project done with Wageningen University & Research (WUR). The model is used to automate visual inspection of tomato seedlings, hence it is crucial to have a highly accurate model.

Read the full story

HOW RETAILERS CAN USE AI TO INCREASE PROMOTION PROFITABILITY BY 20%

OPTIMIZING RETAIL PROMOTIONS BY PREDICTING CUSTOMER LIFETIME VALUE EFFECTS - “Buy one, get one free… Don’t miss this special offer: click here for 2 months premium membership.” Promotions. Retail is hard to imagine without them. Spending on promotions in Europe has doubled over the past 10 years, with nearly 30% of purchases being on promotion. This increased promotional pressure is a result of different factors. Increased competition from e-commerce and changing consumers expectations, to name a few. As such, the ability to design winning promotions with positive sales and margin impact is more important than ever. Unfortunately, many retailers get stuck at basic levels.

Read the full story

From gut feeling to data-driven actions

How does a data value case look like in the perfect world? In that perfect world, it would start with a C-level sponsor who is convinced of the value data can bring. She (or he) can clearly state this value case: the question she has and how she thinks data can answer this. She realizes that this will not happen overnight, but will take effort, time, and attention. And of course, all of the data that is needed to bring her vision of value to life is fully and easily available, clean, and hardly needs any processing. So, bring out the magic wand and transfer the data into useful day-to-day value.

Read the full story

On satellites, growth curves and the birth of the Crop Detector

Dutch people are pretty level-headed. You don’t hear Dutch people say that we are the greatest country on earth, or that we have the best food. But the Dutch sometimes forget to be proud of what they have accomplished. After years of battling the sea, specifically the ‘Zuiderzee’, we had enough and turned this sea into a lake. But maybe the biggest accomplishment of all is that we have created our own land: Flevoland. Since 1950 we have built an entire extra province. After having suffered many food shortages during war periods, Flevoland was constructed to secure food supply. Flevoland turned out to host the most fertile soil of the Netherlands. These days, about four times as much crops are grown per square meter in Flevoland compared to the rest of the Netherlands. A lot of these super-fertile fields are owned by the government, and leased to farmers. In order to ensure that the quality of the soil remains high and to avoid diseases in the ground, a crop rotation plan needs to be followed, see Figure 1. Crop rotation plans are therefore imposed on the farmers by the government.

Read the full story

Hackathon on sales prediction modeling

Our Itility September hackathon was based on creating a sales prediction model for beer. A real customer case – with the goal to predict the effect of promotional activities on the number of crates sold. The goal was to predict the sales for week 34 of 2019 in the provided data set. A combination of clustering, regression and some common sense did the trick for the winning team.

Read the full story

Object Detection Metrics

Anyone concerning themselves with object detection sooner or later encounters this problem: you have created your first object detector and now you want to know its performance. Sure enough, you can see the model finds all the objects in the pictures you feed it. Great! But how do you quantify that? You are familiar with precision and recall and might even have seen some precision-recall curves in the paper you just read, but how exactly do you get your model to produce those curves? And what is that magic number – “AP” – that they put in table 1? Fear no more, we will get through this together.

Read the full story

Outperforming the human eye with a deep learning data generation system

In earlier blog posts on the next-gen tracking solution for football, BallJames, we have looked into several deep learning components that make up our solution: the YOLO-based detection algorithm, our tracking component and the ball tracking system. This post talks about yet another component: team and player identification. Crucial components, as they drive the ability of BallJames to correctly decide which player is detected without any human assistance.

Read the full story

Back on track: locating assets on aerial images

Before the summer, the Geronimo.AI team has completed a six week project for ProRail, the Dutch organisation that takes care of the maintenance of the railway infrastructure. The actual maintenance of rail sections is however tendered to subcontractors. It is important in this tendering process that it is known which and how many assets are present in a certain rail section. If mistakes are made, ProRail will be penalised. The aim of the project was to lower the risk of penalties.

Read the full story

Improving clustering with adversarial training

Current AI-based clustering approaches face challenges when they’re being used on complex data such as images. Developing a solution to this issue would have large practical implications as it could further propel AI’s capabilities in a variety of areas. This blog is the first in a three part series summarising some recent research, aiming to better understand the black-box models used for deep-learning based clustering and to further improve it.

Read the full story

Development of a data-driven road maintenance tool

Currently, road maintenance in the Province Zuid Holland is performed periodically: every 6 or 12 years maintenance is performed. Together with the Province we, Geronimo.AI, are exploring condition-dependent maintenance. The core challenge is to develop a predictive model for road quality which we are developing based on historic and continues datasources and are elaborating in this post.

Read the full story

Building chatbots: Be patient and grow your assistant

For a long time, chatbots have been the number one example of putting artificial intelligence (AI) into practice. Although there are much more sophisticated AI applications, the simple chatbot interfaces make the technology tangible and accessible to the average user through smartphones and messenger apps. However, a massive breakthrough of chatbots is still to come. So far, interactions remain too linear to simulate the feeling of a fully fledged digital conversation partner. What’s on the outlook for chatbots? And what challenges do you face as developer when building one? We discuss this with Tim Groot, AI developer at e-office and speaker at our latest Developer Night.

Read the full story

Hackathon on prediction models for car-sharing with Amber

In the first hour of the hackathon, most teams chose to split up, part of the team exploring the data and the other part trying to find external data sources. There were some splendid ideas, like including traffic information and demographic information of the hub region. Tragically though, in the time that followed, most of the ideas seemed infeasible.

Read the full story