Page 123 - READING COMPANION TO ENGLISH EXAMS

7.2 ROBOTICS AND ARTIFICIAL INTELLIGENCE

From Clay to Code:

Translating the Tablets of Ancient Mesopotamia

The history of ancient Mesopotamia is captured in clay tablets inscribed in cuneiform, documenting its political,
social, economic, and scientific aspects. However, these records remain substantially untranslated owing to
their large number and the scarcity of qualified experts who can crack the code. To eliminate these hindrances,
a team of Assyriologists, historians of the ancient Near East, did their utmost to develop a machine-learning
model. They created two versions: one using the Latin script representation of cuneiform signs and the other
using the digital representation of characters. Based on the study results, the former delivered more satisfactory
outcomes, achieving a score of 37.47 in BLEU4, an algorithm that meticulously assesses the similarity level
between machine and human translations. The program works best for sentences with 118 characters or fewer.
In some sentences, it produced ‘hallucinations’—output that was grammatically correct but stood in need of
accuracy. The developers state the program’s translations can be utilised as preliminary text analysis. They
suggest integrating machine translation as part of a human-machine collaboration strategy where scholars can
edit and polish up the model’s output. The team came to the conclusion that translation is a fundamental activity
with a long history dating from the inception of writing, calling for expertise in two languages and cultures.
Despite the growing accessibility of digital translation tools, ancient languages remain challenging due to the
need to understand long-dead linguistic communities and often fragmentary, incomplete texts. Advances in
optical character recognition and machine translation have improved, but challenges hang in the air.

a Match the highlighted words and phrases in the passage to the definitions below.
1. : information or images that are produced by a computer, machine, or printer
2. : to achieve or produce something that has been promised
3. : made of small parts that are not connected or whole
4. : preceding or done in preparation for something fuller or more important
5. : something that delays or prevents progress by making a situation difficult
6. : to write or carve words or symbols on something, especially as a formal or permanent
record
7. : to work on improving something
8. : by a large amount or degree; very much
9. : in a way that shows great attention to every detail
10. : to need or deserve a particular action, remark, or quality

b Replace the words or phrases in bold with one of the underlined phrases in the passage.
Change the form if necessary.
1. As a result of long-term research, a team of scientists has deduced that machine-learning algorithms can
help identify anomalies and patterns in large data sets more efficiently than traditional statistical methods.

2. While artificial intelligence has made significant progress in recent years, it still lacks the ability to replicate
human creativity and imagination.

3. The decision-making processes of artificial intelligence algorithms are sometimes difficult to decipher even
for their creators, which is known as the ‘black box’ problem.

4. Despite significant advancements in the field, several ethical and technical challenges surrounding the
explainability and bias of artificial intelligence models remain unsolved.

5. Robotics engineers try hard to design and program robots with advanced artificial intelligence and
machine learning algorithms that enable performing complex tasks with precision and accuracy.

122
Ortaöğretim Genel Müdürlüğü Öğretim Programları ve Ders Kitapları Daire Başkanlığı