Artificial Intelligence

The software of the future is intelligent. We help to evaluate and efficiently use the methods and technologies from the research field Artificial Intelligence (AI).

On one hand, this discipline has been around for more than 60 years. On the other, it has recently received a tremendous boost from such new technologies as big data and cloud computing. For inovex, as IT project experts with an integrated digital transformation portfolio, AI is a particularly exciting field, as it has many overlaps with our other areas of specialisation.

What is Artificial Intelligence?

Artificial intelligence is a large field which encompasses all those digital systems designed to behave like human beings. These systems are, however, capable of performing only specific, narrowly defined tasks, rather than genuinely comprehending and empathising with the full spectrum of human behaviour. If a system can perceive excerpts of reality, evaluate options, make decisions, and trigger meaningful actions like a human being, it is an AI system.

At inovex, our AI research is focused on three core areas: processing visual input using computer vision, facilitating computer-aided communication using natural language processing, and bundling application areas in robotics.

Interdisciplinary Team of Experts

inovex has built up a large team of proven artificial intelligence experts. More than 100 inovexperts are currently working in data management & analytics, a field which also includes the underlying and associated technologies and methods. Our extensive, in-depth knowledge enables us to convert demanding and complex requirements from these disciplines into productive solutions.

Technologie-Stack
  • TensorFlow
  • scikit-learn
  • OpenCV
  • DLIB
  • scikit-image
  • Keras
  • Spark
  • Apache Tika
  • MITIE NLP
  • nutch
  • spaCy
  • PyTorch
  • transformer
  • Rasa
  • huggingface
  • sentence-bert
  • MLFlow
  • DVC
  • kubeflow
  • azure ML
  • Chatterbox
  • Lex
  • Kubernetes
  • langchain
  • haystack

Generative AI

EXPLORATORY WORK WITH AI

Generative AI has recently made a name for itself with AI-driven image generation via DALL-E and text generation via GPT-3. We take a wide-ranging look at the opportunities that Generative AI holds for enterprises. We take an exploratory approach that starts with feasibility studies, continues with product discovery and proof of concepts, and finally leads to a new project.

In an internal research project, we are working on the cutting edge of language models. Here we continue to build our technical understanding and work on the effective use of the large language models. We are happy to pass on the knowledge we have generated in this way in the form of workshops.

Prompt Engineering

Customized prompt solutions

Conversational AI such as chatbots or digital assistants, text and document understanding as well as automated content creation: Prompt Engineering enables the effective and targeted control of generative AI models. We help you to identify potential use cases for your company and implement them using generative AI models.

Computer Vision

RECOGNISING AND UNDERSTANDING IMAGE CONTENT

Our AI offerings are closely linked to the machine-based algorithmic processing of sensory input, i.e. the perception, analysis, and further processing of spoken and written language, as well as of images and videos. Examples of use cases include scanning and interpreting vehicle registration documents, POS receipts, and similar items in order to improve the user experience.

In the field of self-driving vehicles, AI solutions can use videos and images to detect obstacles. Initial approaches have also been made in using neural networks to predict whether or not a pedestrian intends to cross the street.

Computer vision is also used in industrial processes. During automatic optical inspections, it enables the fast, consistent, and complete monitoring of industrial products – around the clock. In this context, computer vision does not necessarily operate with ordinary coloured images. Instead, it can be combined with any imaging technique. Multispectral cameras, for example, can be used to reveal the chemical properties of materials, while x-ray sensors can be used to see inside objects and 3D sensors can be used to identify their shapes.

Natural Language Processing (NLP)

SUCCESSFUL COMMUNICATION IN TEXT AND SPEECH

Natural Language Processing enables the automated resolution of text- and speech-based problems. These range from seemingly simple tasks like predicting a missing word in a cloze text to complex challenges like translating or summarising texts.

Typical customer questions include “How can key information be filtered out of long texts – such as contracts – without someone having to read the text?” and “How can text information be automatically visualised?”

Our services range from implementing proofs of concept to developing productive systems. inovex experts are currently intensively researching the possibilities offered by using neural networks to process texts in order to handle linguistically complex customer data records more efficiently.

Robotics

MANIPULATION AND MOVEMENT

Other typical uses for artificial intelligence include evaluating social media posts and performing sentiment analyses. It is also deployed in AI-based data products such as recommender systems, voice assistants, and voice APIs — and even in robotic arms and robots. We are, for example, working on human-like robotics solutions such as Pepper and Nao, which are able to interact with people and are used, for example, in retail environments.

Our Offering

AI and Machine Learning in Production

In industry, especially in production, too much data is produced to be processed by humans. AI solutions can be used to intelligently process these huge data volumes and to derive knowledge from them. inovex has been successfully implementing this data science for manufacturing companies for many years. We will ensure that your machine learning models are implemented reliably, reproducibly, stably and scalably.

Developing Algorithms in Python, Scala und Java

In order to facilitate machine learning, algorithms must be developed. Our experts combine mathematical algorithmic approaches with professional software development.

AI and Machine Learning Research

inovex maintains very close contacts with universities and research institutions. We are constantly working on the very latest research topics, enabling our senior data scientists to support you in all your projects. Whether you want to set up an AI lab or develop advanced models such as abstract text summarisation, segmentation, keyword extraction and automated visual inspection models, we will give you an overview of the current level of research and technology, perform a feasibility study, and implement your ideas.

AI in Industry

Artificial intelligence can be used in industrial settings to quickly and reliably identify problems and weak points in machinery. We will help you set up predictive maintenance solutions and automate your quality assurance processes (predictive quality). Contact us to see how we can assist you.

Deep Learning & AI Consulting – Model Clinic

Developing reproducible machine-learning models requires both high-quality code and a considerable amount of experience. If you have already taken the first steps towards using machine learning but now need support, we will be happy to advise you. Our expertise also includes interpreting and understanding the behaviour of models.

Cloud AI

All leading cloud providers now offer AI services. We will help you to understand and make the most of the opportunities offered by MLaaS (Machine Learning as a Service) APIs.

 

AI Product Discovery Workshops and Design Sprints

Many companies first want to understand where in their value chain AI can be used to give them a head start. In order to help companies identify these points in a structured manner and to determine the initial starting points for solution concepts, we offer special AI Product Discovery workshops, which are aimed at either Product Managers or specialist experts and data scientists. We can work with you to use an AI product prototype and Design Sprints to collect customer feedback and gain tremendous insights in just five days. This process enables weeks’ and months’ worth of progress to be achieved in a matter of days, thus enabling a far-reaching understanding of any problems by gathering direct feedback from internal customers in the cross-functional team.

AI Proof of Concepts and Prototypes

Do you have a specific business problem for which you want to develop an AI solution? We will be happy to evaluate your ideas by carrying out a proof of concept assessment using your data.

 

Trainings & Coaching

For years now, inovex has been offering Apache Spark training courses containing machine learning modules. We have now teamed up with consulting experts from our partner Königsweg to include Python training courses. Upon request, we can also design practical in-house training courses for your company on AI, machine learning or deep learning with PyTorch, Tensorflow or NLP.

 

Case Studies

Artificial Intelligence

mobile.de: Use Cases for Online Portal Recommendations Using Data Products

mobile.de is a marketplace for the buying and selling of vehicles. Every month, the web platform draws 13.5 million visitors who can choose from the more than 1.6 million vehicles on offer. Each visit to the platform creates a stream of data which contains information about the demand for particular vehicles, the quality of the vehicles for sale, and user requirements. mobile.de wants to use this data to continuously improve the user experience for both vehicle sellers and purchasers.

READ CASE STUDY

H-Hotels Gruppe: Platform search and automatic content linking

Making quick and easy room reservations through the H-Hotels website is one of the key digital assets for the company’s business model. Together with inovex, H-Hotels have now set themselves the goal of automatically making the existing website content searchable using the latest search technologies in order to further enhance the platform’s usability for customers.

READ CASE STUDY

beyerdynamic: Headphone Production 4.0

beyerdynamic has stood for high-end headphones, microphones and conference systems since 1924. The products manufactured by the company, which is headquartered in Heilbronn, Germany, deliver audio technology with outstanding sound quality. As a leading audio brand, beyerdynamic integrates innovation into both its audio-technology products themselves and its production processes. It was for one of their innovative production projects that beyerdynamic leveraged inovex’s expertise.

Get in touch!

Florian Wilhelm

Head of Data Science, Contact for Data Management & Analytics