Our Achievements and Future Goals

Simbo.AI has been recognized as the winner of NASSCOM CoE’s Healthcare Innovation Challenge (HIC) in the Prescription Digitization Segment. Recently, Simbo.ai introduced its flagship cloud-based application programming interface (API) platform, SimboAlpha, which empowers healthcare professionals to create voice-based Electronic Medical Records (EMR).

Simbo.ai’s flagship product is entirely hands-free, enabling doctors and medical practitioners to generate medical records simply by dictating them. Its advanced speech-to-text technology converts spoken words into text seamlessly. Current approaches to Natural Language Understanding (NLU) typically focus on AI models trained for intent classification, but they often limit the system’s ability to understand language in a human-like way.

Universal NLU is an innovative method designed to comprehend human language in a more natural manner. It takes spoken utterances as input and produces Human Thought Representations as output. If the spoken content falls within the domain of Universal NLU, it will generate detailed thoughts; otherwise, it creates more basic representations. Universal NLU continuously gathers information for further processing and distinguishes between language-specific syntax and semantic meaning, making it adaptable to any spoken language.

Natural Language Processing (NLP) is primarily conducted through intent classification models applied to sentences. While segmenting sentences in written text is relatively straightforward, spoken language tends to be a continuous flow of words produced by speech-to-text engines, making the extraction of intents from streaming audio quite complex.

Current solutions often require users to adapt their behavior, such as inserting pauses or using wake words, which can be effective but may feel unnatural, especially in settings where patients and doctors are engaged in fluid conversations. Existing models typically rely on punctuation-rich text generated by Speech-to-Text systems, which depend heavily on pauses and language structure. Our research is now focused on a hybrid approach that incorporates pauses, meaning accumulation, and centom theory.

GIPCA (General Intelligence Predictive and Corrective Architecture) is a proprietary framework that serves as a fundamental component of BISLU (Brain-Inspired Spoken Language Understanding). Its name reflects its dual capability to not only predict outputs but also correct them. By employing both supervised and unsupervised learning techniques, it operates successfully with considerably smaller training datasets. BISLU is designed as a cohesive Spoken Language Understanding framework utilizing general intelligence algorithms that extend beyond traditional AI/ML, offering a versatile language comprehension model. It integrates approaches such as Artificial Neural Networks (ANN), Symbolic AI, Knowledge Bases, Predictive Memories, and various other brain-inspired algorithms to accomplish this goal.