Using AI to Discover Intelligent Design
Human senses are excellent design detectors, but sometimes they need a little help. In a recent case, AI tools were applied to aerial photographs of the Nazca plain in Peru. The algorithms, trained on known geoglyphs, were able to select hundreds of candidate sites with figures too faint for the human eye. Many of them, on closer inspection, turned out to indeed contain patterns on the ground indicative of purposeful manipulation by indigenous tribes that lived in the area long ago.
Here is a case where humans used their intelligent design to create intelligently designed “machine intelligences” capable of detecting intelligent design. Even so, the scientists needed to use their innate design detection abilities to follow up on the AI results to validate the potential design detections. AI is a tool, not a thinker. As a tool, it offers new powers to archaeology: one of the examples of intelligent design in action in science.
The Nazca Pampa is designated a World Heritage Site by UNESCO because of its immense geoglyphs, averaging 90m in length. The well-known ones, consisting of lines, geometric figures and images of animals, were rediscovered in the early 20th century and have fascinated scientists and laypeople alike. UNESCO describes what makes them unique:
They are located in the desert plains of the basin river of Rio Grande de Nasca, the archaeological site covers an area of approximately 75,358.47 Ha where for nearly 2,000 uninterrupted years, the region’s ancient inhabitants drew on the arid ground a great variety of thousands of large scale zoomorphic and anthropomorphic figures and lines or sweeps with outstanding geometric precision, transforming the vast land into a highly symbolic, ritual and social cultural landscape that remains until today. They represent a remarkable manifestation of a common religion and social homogeneity that lasted a considerable period of time.
They are the most outstanding group of geoglyphs anywhere in the world and are unmatched in its extent, magnitude, quantity, size, diversity and ancient tradition to any similar work in the world. The concentration and juxtaposition of the lines, as well as their cultural continuity, demonstrate that this was an important and long-lasting activity, lasting approximately one thousand years.
Based on pottery fragments, the geoglyphs are dated to between at least 100 BC and possibly up to the 15th century. The spellings (Nasca vs Nazca) appear to be interchangeable. Mysteries remain about the purpose of geoglyphs, and various theories are debated. One thing is indisputable: they were designed by intelligent minds. The people made considerable effort to modify the landscape for whatever purposes that drove them. But that’s OK; ID theory can detect design without knowing the identity of the designer(s) or why they did their work. ID’s job is done when the Design Filter has ruled out chance and natural law to conclude something is the product of a designing intelligence. Discerning the purposes of designs like these are left in the capable hands of anthropologists, historians, and archaeologists, who may find themselves puzzled by some of the discoveries like the “knife-wielding killer whale” figure.
The New AI-Directed Discoveries
New detections of Nazca geoglyphs have continued slowly through the years. A team of Japanese, European, and American researchers, Sakai et al., publishing in PNAS, boasts that AI has accelerated the pace of new discoveries:
The rate of discovery of new figurative Nazca geoglyphs has been historically on the order of 1.5 a y (from 1940s to 2000s). It has accelerated due to the availability of remotely sensed high-resolution imagery to 18.7/y from 2004 to 2020. Our current work represents another 16-fold acceleration (303 new figurative geoglyphs during the 2022/23 season of field work) using big geospatial data technologies and data mining with the aid of AI. Thus, AI may be at the brink of ushering in a revolution in archaeological discoveries like the revolution aerial imaging has had on the field.
The Nazca geoglyphs can be classified as line-type (carved into the ground) or relief-type (made by aligning stones above ground). They can also be distinguished by subject matter and size. Sakai et al. surveyed the entire Nazca Pampa (629 km2), then subdivided aerial photographs with 10-cm resolution into grids. They trained their AI model on 406 relief-type glyphs and gave the AI some puzzles to solve:
To leverage the limited number of known relief-type geoglyphs, and to render the training robust, data augmentation is paramount. Hand-labeled outlines of known geoglyphs serve to pick 10 random crops from within each of the known geoglyphs. These are also randomly rotated, horizontally flipped, and color jittered. Similarly, 25 negative training images are randomly cropped from the area surrounding each known geoglyph. We set the ratio of positive to negative training images to 10:25 for a reasonable balance between precision and recall.
This method yielded 1,309 hotspots of likely geoglyphs, which the scientists classed as Rank I, II, or III from most to least likely. “Of the 303 newly discovered figurative geoglyphs,” the paper says, “178 were individually suggested by the AI and 125 were not individually AI-suggested.” It still required 2,640 labor hours of follow-up on foot and with drones to validate the AI selections. Nevertheless, this effort represented a quantum leap in design detection of glyphs with such low contrast they were barely visible to the unaided human eye.
New Scientist included photos of some of the new geoglyphs outlined for clarity. The new ones tend to be smaller and located near trails rather than larger roads, leading the scientists to surmise that they were intended for viewing by local groups instead of for community-wide religious rituals. Reporter Jeremy Hsu wrote about the need for human intelligence to corroborate the selections made by AI:
The researchers followed up on the AI suggestions and discovered a total of 303 figurative geoglyphs during field surveys in 2022 and 2023. Of these figures, 178 geoglyphs were individually identified by the AI. Another 66 were not directly pinpointed, but the researchers found them within a group of geoglyphs the AI had highlighted.
“The AI-based analysis of remote sensing data is a major step forward, since a complete map of the geoglyphs of the Nazca region is still not available,” says Karsten Lambers at Leiden University in the Netherlands. But he also cautioned that “even this new, powerful technology is more likely to find the better visible geoglyphs — the low hanging fruits — than the more difficult ones that are likely still out there”.
The authors believe that many more geoglyphs remain to be discovered in the area. Now that design has been concluded, we may understandably wonder what the people had in mind when they made these figures:
Line-type geoglyphs predominantly depict wildlife-related motifs (e.g., wild animals and plants). Most relief-type geoglyphs (81.6%) depict human motifs or motifs of things modified by humans (33.8% humanoids, 32.9% decapitated heads, and 14.9% domesticated camelids). These do not appear in the line-type figurative geoglyphs at all. Decapitated heads are sometimes depicted alone, while humanoids are repeatedly depicted with decapitated heads and together with domesticated camelids. Examples of both are shown as Insets to Fig. 5. Wild animals, which dominate the line-type geoglyphs, represent only 6.9% (47 geoglyphs) of the relief-type geoglyphs. These include bird, cat, snake, monkey, fox, killer whale, and fish.
Again, though, figuring out the meaning of the designs is not ID’s job. ID is equally valid at detecting evil designs and good designs. Undoubtedly future archaeologists might have trouble understanding 21st century graffiti if they happened upon a destroyed U.S. city without written records or history. But thanks to the Design Filter, determining whether contemporary “art” was designed or not would be a straightforward project
No comments:
Post a Comment