Cognitive Search: A Practical Application of Machine Learning
TECHNOLOGY

Cognitive Search: A Practical Application of Machine Learning

Z S
Zohar Strinka
Wednesday July 05, 2017

In optimization there is an enormous focus on finding the optimal solution. While we do sometimes design heuristics (algorithms that seem to just generally do well), the primary goal is to come up with algorithms that will consistently give you the best solution. The optimal solution is important mathematically, but it is the predictability of the algorithm which is often most important. We know that given the assumptions of the model, we will not lose anything extra by asking the computer to finish solving the problem.

Machine learning on the other hand is a whole different paradigm. With enough math expertise you can understand how a neural net is transforming data into classifications, but it is a much fuzzier relationship. It is unclear how a change to the training data might impact the effectiveness of the model. The choice of certain parameters also involves almost more art than science in an attempt to get a model with useful outputs.

Despite these challenges, there is already a huge benefit for many who are implementing machine learning tools. An experienced data scientist can effectively choose the right algorithm, training data, and parameters to provide a tool which processes data in a way humans simply cannot. Even without that expertise, the tools have gotten so accessible that anyone who can program can create a bad model. But for all the successful machine learning projects, there are often major challenges to building these tools.

And that challenge brings us to the title of this post. “Cognitive Search” is an approach to processing data to provide the user with the right information to make better decisions. Some key technologies that are proving successful in this domain are Sinequa, HPE IDOL and Attivio.

In practice, these solutions involve training machine learning algorithms both on the way people ask for information and on the business-specific information they need. It then becomes relatively easy for a person to understand what the machine learning algorithm is doing, while providing the value of searching for information in a way humans cannot. In short, cognitive search supports decision making in a natural way.

In support of Mashey’s mission to provide companies with easier, faster and more comprehensive access to information for decision-making, Mashey has partnered with Sinequa to provide implementation services and support for their customers.

Contact us to see what Cognitive Search can do for your company.

hello@mashey.com

Mashey logo
Mashey is a next generation data and analytics consultancy that designs and implements modern strategies that help transform companies into data-driven and data-informed teams. Through business intelligence, data warehousing, and data science, Mashey provides companies with the advantages of veteran experience and startup agility.