Alham Fikri Aji

Adjunct Assistant Professor, Data Science

E: alham.fikri@monash.edu

alham-fikri-aji

Dr. Alham Fikri Aji is an Adjunct Assistant Professor at Monash University, Indonesia. He is also an Assistant Professor at Mohamed bin Zayed University of Artificial Intelligence (MBZUAI). He received his Ph.D. and MSc. from the University of Edinburgh. Aji has worked as an NLP scientist at an Indonesian start-up, then moved to Amazon Alexa, before returning to academia.

Aji's work revolves around efficient and accessible NLP. Specifically, he explores how to efficiently train and deploy NLP models through efficient fine-tuning, model compression, and distillation. Additionally, the lack of data makes many NLP technologies inaccessible, so part of Aji's work involves constructing multilingual NLP corpora for training and evaluation, especially for Indonesian languages. He has also worked on evaluating LLMs on cultural and local nuances, as many NLP systems and data do not capture local nuances. Aji is active in grassroots NLP communities, especially for Indonesian and Southeast Asian NLP communities. He actively leads and works on several community-based research projects.

Aji was active in competitive programming and was one of Indonesia's representatives at the IOI 2010, where he achieved a silver medal. He also participated in and won several ACM-ICPC competitions and participated in the World Finals in 2014.

More information

For more up-to-date publication list, you can visit Aji’s Google scholar page.

Research

Efficient NLP, such as:

  • Knowledge Distillation
  • Parameter Efficient Finetuning
  • Mixture of Experts
  • Model compression

Multilingual/Indonesian NLP, such as:

  • Corpus building
  • Synthetic data
  • Zero-shot crosslingual generalization
  • Locally and culturally nuanced evaluation in NLP
  • Code-switching and Code-mixing

Teaching

  • Introduction to Data Science (Monash University, Indonesia)
  • Advanced Natural Language Processing (MBZUAI)
  • Deep Learning for Language Processing (MBZUAI)
  • Exploring the Use of Generative AI for Moderating Online Polarisation (PI), Monash Data Futures Institute, Amount: AUD $49,500 (2023-2024)
  • Inclusive Language Technologies: Data Curation and Large Pretrained Language Models (PI), Monash University FIT, Action Lab, Amount: AUD $33k (2023-2024)
  • Social Media Analysis of Misinformation and Vaccine Hesitancy in Three Middle-Income Countries (Co-PI), Center for Emerging Infectious
  • Diseases Policy & Research (CEID), Boston University, Amount: USD $25k (2022-2023)
  • Exploring the evolution of racial biases over time through framing analysis (PI), Google Research Scholar Award, Amount: USD $60k (2021-2022)
  • Explore CS Research ( Research Workshop for Female Undergraduates) (PI), Google, Amount: USD $18k (2020)
  • LEARN : Label-Efficient Active Resilient Network (Co-PI), DARPA Learning with Less Labels (LwLL), DARPA, Amount: USD $462k (2019–2022)
  • Semi-supervised Learning of Multimodal Representations (Co-PI), DARPA Active Interpretation of Disparate Alternatives (AIDA), Amount: USD $200k (2019–2020)
  • Bridging Linguistic and Visual Knowledge through Visual Genome (PI), BU Hariri Institute Research Incubation Award, Amount: USD $28k (2019–2020)
  • BIGDATA: IA: Multiplatform, Multilingual, and Multimodal Tools for Analyzing Public Communication in over 100 Languages (Co-PI), NSF, Amount: USD $1m (2018–2022)
  • Best Resource Paper Award, EACL 2024
  • Best Resource Paper Award, AACL 2023
  • Outstanding Paper Award, EACL 2023
  • Outstanding Contribution Award, WNGT 2019
  • World Finalists, ACM-ICPC 2014
  • Silver Medalists, International Olympiad of Informatics (IOI) 2010