PhD opportunity in collaboration with Athens Univ. of Economics and Business


We are pleased to announce a PhD opportunity in collaboration with Ion Androutsopoulos Department of Informatics, Athens University of Economics and Business, under the auspices of ARCHIMEDES Research on AI, Data Science, and Algorithms.

Towards finance market prediction using fine-grained evaluation of LLMs.

Leveraging LLMs in the financial domain presents numerous advantages ranging from data analysis and extraction, risk assessment, stock market prediction, fraud detection, regulatory compliance and customer support, among others. Integrating LLMs in the financial domain enhances efficiency and scalability of processes. An example of a finance market prediction task is stock market prediction. Its performance depends on a subtle interplay of dimensions such as sentiment, emotions, stance, argumentation, uncertainty and temporality. To improve model performance, fine-grained evaluation approaches which are able to explore not only the interplay between dimensions, but also different prompting strategies, domain specific and common-sense knowledge, will be examined. This project aims to refine the integration of LLMs in market prediction and trading by leveraging PIXIU, an advanced framework tailored for the financial domain. PIXIU encompasses a suite of specialized financial LLMs, a versatile multi-task and multi-modal instruction dataset, and a comprehensive evaluation benchmark. This involves examining different prompting strategies and integrating both domain-specific knowledge and broader common-sense information to improve predictions and trading strategies. The goal of the PhD is to achieve higher efficiency and scalability in market prediction and trading processes, ultimately leading to more informed decision-making and better outcomes in the financial sector.

Entry requirements: First-class bachelors or MSc in Computer Science or international equivalent.

Skills: Deep Learning for NLP; prompting, fine-tuning LLMs, RAG, and other current methodologies for generative AI and LLMs; excellent coding skills, ability to analyse large volumes of data, perform statistical analyses, and assess the effectiveness of interventions.

Publications: One or two publications as first author in top-tier NLP/ML/AI conferences (ACL, EMNLP, NAACL, AAAI, NeurIPS)

The successful candidate will have the opportunity to spend time at the National Centre for Text Mining, University of Manchester

Interested candidates should send their CV to The position will remain open until filled

Previous itemNext item
Back to news summary page