Data continues to be a trending topic in all economic sectors and extracting value from it is becoming an increasing priority. This has led to an exponential demand for training in soft and hard skills (functional and technical knowledge) in relation to concepts such as Big Data, Cloud Platforms, Architectures in cloud projects, process automation, data science or BI among others. But especially in recent years, there has been a growing application of AI and ML techniques in different sectors.

In this article, we intend to review some of the most widespread applications during 2021 related to data projects, as well as the main factors involved in solving the potential problems that companies immersed in this digital transformation will encounter.  

Because of, or thanks to?

Over the last few years, many companies have found necessary to migrate their on-premises services to any of the current public cloud platforms. This has allowed them to accelerate the development of use cases based on insights extracted from their data using the aforementioned techniques.

Terms such as forecasting, predictive maintenance, clustering or planning are now the everyday life of these companies. We have also learned to use social networks to improve our results with targeted campaigns, obtaining higher conversion rates.

The future is now

The process of digitization of industry is a reality, and with the implementation of multiple IoT devices that generate huge amounts of information, the real challenge is to transform it into actions that help prevent, alert or advise with a high degree of success. An example of this can be seen in the early diagnosis of diseases or in the prototypes of self-piloted vehicles. And speaking of security, we cannot overlook the enormous investment made in this sector to mitigate the vulnerability of systems and give confidence to the implementation of new technological projects.

Strategic technology trends during 2021

During this year we have seen an acceleration in companies for the incorporation of strategies for extracting value from data that has had an impact on multiple projects in various areas such as:

  • Business forecasting: enables prediction of business development in changing conditions to improve decision making by contemplating real-time forecasting scenarios and having a great impact in cases such as intelligent risk management.
  • ML automation: the maturity of cloud platforms is reducing time to market, enabling end-to-end automation and mitigating manual actions.
  • Effective marketing: presence in social networks, use of SEO, email personalization or sentiment analysis is enabling many companies to estimate churn rate, personalized recommendations or market segmentation.
  • IoT evolution: growing trend in cyber-attacks can affect the prestige of many companies and cause critical economic damage. In this scenario, zero trust models, remote work protection or information leakage prevention have been the workhorses.
  • Customer data capture: a company’s most valuable data are those of its customers. The protection of personal data is being given priority attention in IA projects.

What about tomorrow?

Cloud platforms are increasingly aligned with these needs for scalability, security, connectivity and service integration. One of the clear examples is the integration of cognitive services with other types of services such as analytical exploration. 

These platforms are naturally connected, allowing complex problems to be solved with modular architectures that can be applied both in real time and in deferred time.

Thus, structured or unstructured information (voice, image or text) can be processed with a high degree of effectiveness by these cognitive services available to anyone.

Human-machine interaction has always been one of the biggest challenges in AI, and one of the most popular use cases continues to be conversational assistants. Today we see them everywhere. The clear evolution of these assistants is their collaboration, allowing different bots with expertise in certain areas of knowledge to cooperate with each other and obtain more versatile tools.

Communication will always be an important factor where today there are algorithms that perform simultaneous translations of text into multiple languages or obtain subtitles in real time. The mass production of simultaneous translation devices that allow us to hold conversations in which participants can speak in different languages is on the horizon.

The democratization of AI or ML also continues with firm steps, and this can be seen in the availability of more and more tools that allow us to create architectures and solutions, without the need of having a great technical knowledge. Thus, many simple chatbots can be built from some drag and drops or other tools for data exploration and manipulation in massive data lake ingestions are also starting to be commonly used following the same philosophy.

Last but not least is the research being done to make IoT devices more and more independent and allow greater personalization. Different alternatives are being sought to reduce ML models that can be run on such devices and see how this specific learning can be leveraged by other centralized models in order to take advantage of these synergies and find more robust models that can again be passed on to the reference devices. 

This is undoubtedly a fascinating prospect that has a long way to go, and where the first steps have already been taken.

Technologies that will accelerate growth in 2022

What’s to come is even more exciting and is of great interest to the AI community and companies that need increasingly sophisticated products. As a result, some of the areas where the greatest impact is envisioned are:

  • AI will determine infrastructure decisions: Cloud providers are becoming more strategic in providing greater integration in cognitive services and processes with increasing computational capabilities.
  • Augmented intelligence: increased interaction between robots and humans where processing of unstructured data will be effectively accelerated
  • Increased awareness of ethics in AI projects: proactivity in handling uncertainty and explainability in models and drive for metrics to ensure fairness
  • Real-time communication: NLP is enabling rapid evolution in NLU and NLG which will enable conversational simultaneous translation. Automatic generation of text or image will enable more personalized content.
  • No-coding Machine Learning: cloud platforms work for natural connection between preprocessing, modeling, monitoring, tuning or deployment tools. This will enable rapid development of complex solutions through drag and drop in end-to-end pipeline implementation.
  • TinyML will enable more autonomy in IoT devices: large-scale data capture together with real-time adaptive model design will allow unlimited customization.

 

Image: Unsplash | @sortino

Author

  • Data Scientist in Keepler Data Tech: "Live full, die empty" defines my state. This becomes my lifestyle taking me out of my comfort zone and driving my voracious learning attitude about different aspects of Data Science. I love learning by teaching and am always open to new challenges that push me further my comprehension."