Text Embeddings App

Using text embeddings for lookbacks. I’m making this app freely available to the Norwegian Petroleum Directorate and UK North Sea Transition Authority along with various NLP outputs for the benefit of the geoscience community. This is as input to the hackathon organised by FORCE led by Peter Bormann

This particular example uses 800 license relinquishment reports in the UK for a text embeddings model. From this we can look at the similarity of any term, in this case ‘dry hole’, to various elements defined by the user. The output, in this case a radar chart, is data driven. It uses the statistical co-occurrence patterns and sentiment in the text. It appears to show ‘reservoir’ and ‘trap’ closest to ‘dry hole’ from a word vector perspective based on the 800 reports.

It’s conceptual application can be widespread. Using patterns in past reports to learn and help inform future decisions. This might relate to reservoirs for CO2 storage, groundwater aquifers, subsurface site risks for wind farms and hydrogen infrastructure, fluid flow and deep permeability for geothermal, novel mineral and rock assemblages for new deposits etc.

Where we see the patterns from text generate something which does not fit our existing mental models, it can make us curious to dig a bit deeper to see what is going on. It could be a data artefact or could lead to new learning events, driven by the data (text).

Used in conjunction with other Natural Language Processing (NLP) techniques (not only statistics), this may challenge potential human biases that exist.

By exploiting large volumes of text, the whole may be greater than the sum of the parts. In other words, Big data is about small patterns. These patterns may only be visible when we stack everything together, creating new knowledge that may not exist explicitly in any single report.

I’ve shared many of the other visualisations involving word vectors over the few months. You can also find them on this blog.

#mineralexploration #ccus #geothermal #renewableenergy #oilandgasexploration #hydrogeology #geotechnicalengineering #geohazards

#energytransition #digitaltransformation #naturallanguageprocessing #textanalytics #digitalinnovation #geosciences #subsurface #osdu