Freud

Freud is our ML research effort to develop algorithms that can assess your state of mind solely by analyzing the entries in your journal.

Freud will help you recognize patterns in your behavior, mood, etc., and act as your mental health and lifestyle coach.

Freud, once ready, will be opt-in only and we'll always aim at full transparency with regards to data privacy.

Background

The future is data-driven. In the physiological domain, we currently see a rapid shift from reactive medicine to digital diagnostics and preventive measures enabled by data collected from wearable devices like the Apple Watch and the Oura Ring.

We expect a similar shift in mental health, from clinical psychology to lifestyle coaching and adjustment, not only for the sake of mental well-being itself but also to increase performance and happiness.

The challenge arises from the complexity and peculiarity of our state of mind and from finding a way to access and evaluate it. We believe that a micro journal empowered by sophisticated data analytics has the potential to provide this access and Freud is an experiment to field test this hypothesis.

Experiments

Ask my Journal

We are currently experimenting with a project by Sahil Lavingia called Ask My Book. It uses Open AI's GPT-3 model and the Embeddings API to answers questions about a specific book. In theory it should be possible to use the same technique to answer questions about the entries in a journal, somewhat like a second memory with superpowers.

In case you are interested in this experiment, please let us know.

Roadmap

F/1 – Sentiment

The first stage of Freud will be simple sentiment analysis on a bullet point level to test feasibility and accuracy.

Status: under development

F/2 – Correlation

Once sentiment analysis is established F/2 will attempt to correlate it with context, i.a. #tags, @mentions, dreams, highlights, and so on.

Status: planning

F/3 – Patterns

If a correlation can be achieved, the logical next step is pattern analysis aka complex correlation between multiple entities, and over time.

Status: not started

F/4 – Analysis

The patterns established in stage 3 act as input for ML algorithms that analyze how these patterns are connected, what their origin is, and how they can be leveraged for mental well-being and performance.

Status: not started

F/5 – Insights

In the last stage, algorithms developed in F/1-4 are converted into actionable insights and concrete lifestyle-enhancing recommendations that will be available in the app.

Status: not started

If you are interested in our research or want to contribute, get in touch at research@journalisticapp.com.