TechnologyCogito. Human-like Understanding for Intelligent Process Automation and Knowledge Discovery
The ability to understand language and extract data in documents, manage interactions in natural language and handle unstructured information is a critical factor for competitive advantage for enterprises in any industry.
Cogito: At the forefront of today’s Artificial Intelligence revolution
Cogito® is a platform based on Artificial Intelligence algorithms that mimics the human ability to understand language and complex textual information at scale and with the accuracy required by today’s business world.
Based on a rich, domain-independent representation of knowledge (Knowledge Graph), Cogito integrates the best of Natural Language Understanding (NLU) and Machine Learning (ML) to intelligently automate business processes and harvest relevant knowledge and critical insights from all the information available in your organization.
Cogito’s Knowledge Graph is a representation of the real world where concepts are defined and connected to one other by semantic relationships. Its coverage of language is both deep and wide. Compared to other text analysis technologies that require cumbersome and resource-intensive initial and constant training, Cogito leverages the embedded knowledge graph together with natural language understanding algorithms to read, comprehend and learn from any text, out of the box.
In addition, it is open, not a black box, and its content and structure can be understood by humans, making it easy to be incrementally adjusted to maximize its performance in any domain.
Natural Language Understanding
Cogito distinguishes the correct meaning of words and expressions in context and automatically associates the attributes of more general terms that are conceptually linked to these words. For example, it understands the terms “SUV” and “sedan,” and it automatically comprehends that these words share similar attributes that derive from both being a kind of “car.”
In addition, Cogito identifies the relationships between concepts in a text. Both features derive from its ability to perform different levels of linguistic analysis (morphological, grammatical and sentence analysis) in conjunction with semantic analysis and word disambiguation. This human-like ability to read text and understand language is an additional differentiation from other text analytics platforms because it increases accuracy, speed and the ability to handle complex text.
As the accurate comprehension of textual content always requires domain-specific knowledge, Cogito’s knowledge graph can be extended and expanded based on the specific requirements of each use case.
Unlike other text analytics platforms, Cogito can add domain-specific knowledge through both subject matter experts and through our proprietary machine learning algorithms that leverage the analysis of a training set of tagged content. The balance between the two different methodologies can be managed and A/B tested to make sure you always reach of the quality requirements of your projects.
Cogito has been successfully implemented by hundreds of organizations across industries for a range of business processes by reliably enabling human-like comprehension and insight at scale. Cogito’s Knowledge Graph can be made more effective and trained to work in specific domains by leveraging a rich set of tools designed for data scientists and leveraging subject matter experts and the automatic enrichment of the knowledge graph through machine learning
Cogito has been extensively deployed in many industries thanks to the accuracy of its AI-based, multilanguage Natural Language Understanding and virtually unlimited linear scalability. Cogito can reach over 90% accuracy in understanding content in completely open or domain independent contexts, such as general news and other forms of general communication. Other platforms cannot be directly compared as they do not provide the same capabilities of understanding language; instead, they start from 45-55% and may reach up to 73-75% for a few systems in closed, domain-specific, contexts.
Cogito’s rich out-of-the-box knowledge graph, its powerful tools and its machine learning algorithms provide a more advanced starting point and a more efficient way to customize the technology to work in specific domains or use cases and to maintain the same level of accuracy when conditions change (new words, new tasks, new requirements). This allows Cogito to offer lower development costs and significantly improves the Total Cost of Ownership compared to legacy technologies.
Cogito provides shorter time-to-deployment than other approaches thanks to our out-of-the-box and open, understandable Knowledge Graph. The powerful features of our Knowledge Graph make it possible to enrich and customize Cogito specific to different domains or tasks in a fraction of the time that other technologies, which are unable to understand the content of a document in context, require.
By making information analysis faster and more accurate with a contextual understanding of all kinds of information, Cogito augments analysts’ capabilities for search, information management and analytics to automate processes, generate actionable insight and drive more reliable decision making.
Cogito qualifies as an Explainable or Transparent AI platform, as the performance of its algorithms can be trusted and easily understood by humans. Unlike black-box AI, you can understand how Cogito arrives at a decision and see the logic by which it “reasons” and why it produces certain results. This is a major advantage because if you know how AI works and why it determines particular outputs, you can incrementally refine it and make sure the actions are aligned with your intended objectives.