Collaborating Authors

Out of the box Elasticsearch


When it comes to managing huge data at an enterprise level, intelligent search tools are required. You may have any amount of data but it makes sense only if it can be found and converted into meaningful information. You need a tool that can organize and index data in such a manner that information can be found on-demand. SearchBlox uses the Elasticsearch API which is intelligent enough to remove dysfunctional nodes and connect through alternate routes so that search is available to you at all times. If you are looking forward to investing in a search tool for your enterprise, you must choose one based on Elasticsearch because of the multiple benefits it will instantly offer.

Practical applications for Machine Learning for multi-source text data


Machine learning is the process by which we teach the machine to try to think like humans by feeding it data that is labelled. Embedding a machine learning or deep learning model for text data involving natural language processing of text as well as building predictive capabilities requires a platform where the data is aggregated, processed, analyzed and insights derived at scale. SearchBlox provides the ability to add a deep learning model for any type of source data. Once setup, the cognitive model can be applied to text information to derive recommendations and insights. Here are the practical applications achieved through the deep learning models within SearchBlox for any type data being indexed from the 75 connectors that SearchBlox provides.

SearchAI SmartSuggest : Better search relevance starts with better search queries. How AI can power your search box?


What is an autocomplete or autosuggest feature in search? Autocomplete or autosuggest uses the partial search terms that a user is typing in to complete the rest of the search query so the user can simply click to search from a drop down box in the search form. What are the challenges with autocomplete/autosuggest? Autocomplete can use previous search queries or use content that has been indexed by the search engine to provide the completion suggestions, however there are several issues which provide a bad search experience: •Completion does not take into account where the search term occurs •No relevance ranking for the suggestions •May show incorrect results when search queries are used for completion •Content based suggestions may not be accurate for large documents/web pages/websites •Manual management of autocomplete is required in most cases How can SearchAI SmartSuggest predict better search queries leading to better search queries? SearchBlox SearchAI uses SmartSuggest to predict search queries that lead to more relevant search results. SearchAI understands the search queries and how they relate to the documents using deep learning based NLP processors that read through the content and understand how they are related.

Ask a question? AI model provides answers from your web pages! Right from your search box.


SearchAI Answers is an Artificial Intelligence as a Service (AIaaS) service from SearchBlox that offers direct answers to natural language search queries using an AI model built using the customers' content, without the need for any manual tagging or a domain specific taxonomy or creating a knowledge graph. Search Engines as Question Answering (QA) Engines will replace the traditional approach of information retrieval where we are presented with a list of search results links. What is driving the transformation from search to answers? • New generation of searchers using Siri, Echo & Google Assistant using voice as a primary channel of getting information • Want to ask direct questions and receive direct answers • Ask questions in natural language or a conversational manner • Need concise, relevant and context based domain specific answers SearchAI Answers use explicit and implicit feedback to continually improve the quality of answers using MLOps. What are the business benefits of using SearchAI Answers on your website or portal? Contact us to get started and we will have you running our AI service very quickly.

What is autosuggest? Why do search boxes need to have autosuggest?


Autosuggest helps users find what they are looking for without having to type-in the whole term, product name or key phrases saving user's time and also increasing the relevance ranking of the result they are looking for. Mobile traffic to site search boxes have made the autosuggest a must have product feature that predicts the right terms or phrases for the user to click on rather than type-in the whole term. Mobile users are prone to entering incorrect spellings as evident from email messages sent from a mobile where it is common to see the phrase "Sent from mobile. If the incorrect result is clicked, or the correct link is not found then we have to go through the navigation links to go through each of the search pages to identify the appropriate document or web page. Using an autosuggest, where users can pick the right phrase makes a significant difference in reducing the number of results and ensuring the right results are shown with relevance.