Natural language processing

An example of NLP is Siri on your iPhone. Your audio is recorded and transcribed for the AI to process the message. Various machine learning models are used to analyze individual components of speech.

Sentiment, meaning, and context are all processed when audio is sent for NLP. The way these work is through algoritihmic prediction but even though language is processed mathematically there are still problems.

Language is latent with ambiguity, sarcasm, and unclear contexts. AI methods can deal with this through diverse datasets and continued human reinforcement in training.

Multimodal communication

Let’s bust a myth! Text and audio is not the only way AI can communicate with humans.

The last wave of innovations that propelled us into this new age of AI that doesn’t seem to be slowing down, namely transformer architecture, the TPU, and LLMs. We are constantly seeing different models of communication.

AI is now making the transition to visual and even physical communication. Gaining access to new layers of communication and thus new insights. Using facial expressions, gestures, and sound.

Combining so many different mediums require the coordination in a way that avoids confusion for the end user.