Erik Arakelyan

|

About Me

I am a Machine Learning Researcher at NVIDIA with interests across all things Reasoning within AI, Explainable AI and Knowledge Graphs. I have completed my Ph.D. at the University of Copenhagen under the guidance of Isabelle Augenstein and Pasquale Minervini. I am a former member of the CopeNLU research group and the Pioneer Centre for Artificial Intelligence in Denmark.

My research spans developing reasoning methods across different modalities in Deep Learning, advancing explainable AI techniques, and enhancing reasoning varifiability and faithfulness in neural models through augemntaion with Symbolic methods. I am also interested in the evaluating the optimality of reasoning methods in neural models and their impact on the model’s performance and generalization.

Among other engineering and research experience, I had visiting research stints at Amazon’s Alexa AI team and Cohere AI, working on grounded, faithful and varifiable LLM reasoning. Additionally, I’ve led model tailoring and optimization teams at ARM as an Apllied Machine Learning Enginer.

Previously, I completed my MSc in Machine Learning at University College London in the Machine Reading Group, where under the guidence of Pasquale Minervini, my thesis, “Complex Query Answering with Neural Link Predictors”, received the Outstanding Paper Award at ICLR 2021. My academic journey began with a B.S. in Computer Science from the American University of Armenia, supported by a full governmental scholarship and the “Best Bachelor in The Sphere of IT award”.

In my free time, I enjoy playing basketball, boardgames and narative based activites, participating in hackathons, and occasionally indulging in classical music, a passion from my earlier years as a music student.

Selected Publications

2025

  1. With Great Backbones Comes Great Adversarial Transferability

    Erik Arakelyan, Karen Hambardzumyan, Davit Papikyan , and others
    arXiv preprint arXiv:2501.12275 (2025)

2024

  1. Semantic Sensitivities and Inconsistent Predictions: Measuring the Fragility of NLI Models

    Erik Arakelyan, Zhaoqi Liu, Isabelle Augenstein
    EACL 2024 (Oral, Outstanding Paper Award) (2024)
  2. SynDARin: Synthesising Datasets for Automated Reasoning in Low-Resource Languages

    Gayane Ghazaryan, Erik Arakelyan, Pasquale Minervini , and others
    In The 31st International Conference on Computational Linguistics (COLING) (2024)
  3. FLARE: Faithful Logic-Aided Reasoning and Exploration

    Erik Arakelyan, Pasquale Minervini, Pat Verga , and others
    arXiv preprint arXiv:2410.11900 (2024)

2023

  1. Adapting Neural Link Predictors for Data-Efficient Complex Query Answering

    Erik Arakelyan, Pasquale Minervini, Daza Daniel , and others
    NeurIPS 2023 (2023)
  2. Topic-Guided Sampling For Data-Efficient Multi-Domain Stance Detection

    Erik Arakelyan, Arnav Arora, Isabelle Augenstein
    In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (ACL 2023 (Oral)) (2023)
  3. Approximate Answering of Graph Queries

    Michael Cochez, Dimitrios Alivanistos, Erik Arakelyan , and others
    Compendium of Neurosymbolic Artificial Intelligence, IOS Press (2023)

2020

  1. Complex Query Answering with Neural Link Predictors

    Erik Arakelyan, Daniel Daza, Pasquale Minervini , and others
    In ICLR 2021 (Oral, Outstanding Paper Award) (2020)

Latest Posts

The posts below are AI-generated placeholders for demonstration