Forensic analysis of linguistic evidence

Forensic analysis of linguistic evidence

Contact: Shunichi Ishihara shunichi.ishihara@anu.edu.au>

Due to the wide availability of mobile devices and online communication tools, their convenience and high anonymity can be ill-used for criminal activities as a medium of malicious communication. In such case, recorded voices and text messages can be deployed to assist the police investigation or often presented as forensic evidence leading to the identification of the wrong doers. The relevant areas of study are called ‘forensic voice comparison’ and ‘forensic text comparison’ in forensic science and linguistics. The demand for the forensic analysis of linguistic evidence has been increasing.

The summer scholar will work as a member of the forensics team and gain experience in analysing speech and textual evidence and interpreting the analysis, by contributing to empirical studies on forensic voice/text comparison and related areas. The summer scholar will also acquire the relevant fundamental knowledge of forensic science. Depending on the student’s skills and research interests, other academics may be involved in supervision. Tasks may be related to the following two areas, depending on whether the scholar is interested in speech or text processing:

  • Application of speech processing techniques; more specifically the sub-band cepstral analysis, to voice evidence and beyond. Specific topics can be:
    • The analysis of different types of recordings and sounds (e.g. non-vocalic sounds, and emotional, contaminated and synthetic speech)
    • The application to linguistic-phonetic studies
    • The application to forensic voice comparison
  • Validations of forensic text comparison systems. Specific topics can be:
    • The comparison of some existing systems in performance under some forensically realistic conditions
    • Forensic text comparison in a language other than English (e.g. Japanese)
    • The assessment of uncertainty in forensic text comparison
    • The use of AI tools: e.g., ChatGPT, in forensic text comparison and related areas

Some topics require good coding skills and knowledge in speech or natural language processing. Towards the end of the program, the scholar is encouraged to formulate a research question to be possibly tackled as an Honours, Masters or PhD project.

Supervisors: One or more of Shunichi Ishihara, Frantz Clermont and possibly Yuko Kinoshita depending on topics

Updated:  27 June 2023/Responsible Officer:  Head of School/Page Contact:  CASS Marketing & Communications