Hardel’s project managers ensure compliant high-quality deliverables through multi-level supervision, quality control and revision. Our teams can successfully handle delicate projects in the following areas:
Building and management of parallel language corpora
Our team handles the search, collect (internet and other sources), translation/editing and alignment of large parallel corpora.
Hardel has a long experience in the development of highly-annotated lexicons for application in digital dictionaries and machine translation. This experience covers also the maintenance and updating of lexicons (adding special domains, parts of speech, semantic annotations and further meanings).
Translation and management of various translation projects (including special domain projects)
Transcription of audio data for speech projects (voice recognition and other) in different languages. Various transcription tools are used.
Development of normalization tools for different languages and dialects. For example a Language Normalizer for Standard Arabic normalizes/corrects common mistakes, e.g.:
- ىء instead of ئ
- ه instead of ة
While a Language Normalizer for Arabic dialects converts dialect words to MSA, a Language Normalizer for Spanish would restore Spanish special characters (like in “atención” vs. “atencion”) where necessary.
Transliteration involves converting the spelling of words (originally belonging to one language) into a different alphabet. One example is Romanization (e.g., the Arabic name قرقماز is transliterated “Qarqmaz”).
Testing & Feedback
We have a diversified experience in setting up testing procedures to help improve performance of all applications and tools above including quality assurance and control. Our expertise in designing such procedures can be of great help to our customers.
Our team of native speakers contributed to tagging various corpora sets for parts of speech or annotations related to syntax, morphology, semantics, named entities, etc.
Diacritization (تشكيل ) handles adding diacritics (short vowels / حركات ) to the Arabic, Hebrew and other scripts. We contributed to developing tools that apply diacritics automatically.
Hardel has also contributed to several linguistic tools, such as:
Morphology Analyzer – Morphology generator (various languages)
Number converter: converts digit numbers to letter numbers or vice versa
Preprocessing & postprocessing tools
Named entity detection
Part of Speech tagger
Content copyright 2018. HARDELINC.COM. All rights reserved.