Deepmind: do you know about its AI development in archaeology?
The study of archeology symbolizes a key that opens the doors of the past. However, it is still difficult to reveal its secrets. Hence, recent technological revolutions have made it possible to decipher the difficult and deep mind He has been a great ally. Now it’s easier to discover our world by critically exploring those elements of the past!
Archeology in the age of AI
The current challenge in archeology is not to discover new things, but to work on those objects that have already been excavated and digitized. Currently, it is not possible examine all relevant records. So the investigative capacity on the obtained objects is limited.
Along these lines, these are some of the AI advances that help archaeology:
- Discovery of dozens of abandoned settlements long ago along the coast of Madagascar that reveal environmental connections to modern communities.
- Detection of almost imperceptible bumps from the mounds of earth left by the prehistoric cultures of North America.
- Other researchers have cmapped the river systems of the Bronze Age in the Indus Valley, one of the cradles of civilization.
Artificial intelligence supports scientists in the search for new archaeological excavations on a scale never imagined. So it can be said that AI has been a tool for the advances found. The algorithms of machine learning provide a faster path to complex data analysis.
Hence, developments in Artificial Intelligence and Machine Learning have made it possible to classify existing digitized archaeological artifacts, improving data search capabilities and allowing a better understanding of ancient cultures. What does the Deepmind have to do with all this?
Deepmind: Unveiling symbols
DeepMind combines deep learning and algorithms. This through an algorithm that generates an equivalent model that can work with real world data.
According to a study, one of the latest AI models generated by DeepMind helped restore a missing text from ancient greek inscriptions. In addition, he offered suggestions about when the text was written, as well as its possible geographical origins.
It has been pointed out that Google’s DeepMind developed the pythian neural network. Their goal was to fill in the missing ancient Greek inscriptions on the damaged surfaces of stone and ceramic artifacts. Among its particular features are:
- It is named after the Oracle of Delphi.
- Take as input a corrupted text stream.
- Predict character sequence of the hypothetical restorations of the ancient inscriptions.
Deepmind’s newest development: Ithaca
Ithaca is the new software that is based on a dataset of some 78,608 ancient Greek inscriptions. Each one is labeled with metadata referring to where and when it was written. The algorithm looks for patterns in this information, encodes them into complex mathematical models and, using these inferences, suggests:
Going deeper into this, a study reports that the accuracy of this model is 62% when restoring letters in damaged texts. Also:
- Attributes the geographic origins of an inscription to one of the 84 regions of the ancient world with a 71% accuracy.
- It locates the text in an average of 30 years around its known year of writing.
The authors noted that despite a promising outlook, Ithaca cannot operate independently of human experience. The suggestions focus on data collected using traditional archaeological methods. So the architecture of this deep neural network is based on:
- The colaboration.
- Decision support.
- The interpretability.
Solving some controversial cases
East ithaca algorithm had his intervention in a historical controversy around some Athenian decrees and their dating dates. Their date had initially been set at around 446 BC After careful study by historians, some pointed out that the decrees were written around 420 BC Ithaca predicted a date of 421 BC, very close to the previous conclusion (Oulette, 2022) .
As we can see, history provides invaluable information about our past. It makes sense of our current situation and helps us prepare for the future. Researchers increasingly turn to emerging technologiesso there is still a lot to go and learn.