3d-rendering-warp-speed-hyper-loop-with-

Mubashara Akhtar

mubashara.akhtar@ai.ethz.ch

I am currently a postdoctoral researcher at ETH Zurich and a fellow of the ETH AI Center. I am also a visiting researcher at the University of Cambridge, collaborating with Andreas Vlachos.

I completed my PhD at King's College London. I was previously a student researcher at Google Deepmind in Zurich, where I worked with Julian Eisenschlos, and a visiting doctoral researcher at the University of Cambridge.

I am a research chair at the Evaluating Evaluations (EvalEval) coalition. I am also part of the organisation committee of the FEVER workshop and co-lead the Responsible AI committee of the MLCommons project Croissant.

My current research interests include benchmarking & evaluation, (multimodal) reasoning, and VLMs. ETH Master students interested in doing a Master's thesis, a semester project, or getting involved in research collaborations with me, feel free to reach out via email or fill out this form.

Selected Publications

TANQ: An open domain dataset of table answered questions.

Mubashara Akhtar,* Chenxi Pang,* Andreea Marzoca, Yasemin Altun, Julian Martin Eisenschlos

Transactions of the Association for Computational Linguistics (TACL), 2026.

Compose and Fuse: Revisiting the Foundational Bottlenecks in Multimodal Reasoning.

Yucheng Wang, Yifan Hou, Aydin Javadov, Mubashara Akhtar, Mrinmaya Sachan

ICLR, 2026.

When AI Benchmarks Plateau: A Systematic Study of Benchmark Saturation.

Mubashara Akhtar,* Anka Reuel,* Prajna Soni, Sanchit Ahuja, Pawan Ammanamanchi, Ruchit Rawal, Vilém Zouhar, et al.

ArXiv, 2026.

Ev2R: Evaluating Evidence Retrieval in Automated Fact-Checking.

Mubashara Akhtar, Michael Schlichtkrull, Andreas Vlachos

Transactions of the Association for Computational Linguistics (TACL), 2025.

Croissant: A Metadata Format for ML-Ready Datasets.

Mubashara Akhtar*, Omar Benjelloun*, Costanza Conforti*, et al.

Spotlight paper 🌟 at NeurIPS 2024.

Croissant: A Metadata Format for ML-Ready Datasets.

Mubashara Akhtar, Omar Benjelloun, Costanza Conforti, et al.

DEEM workshop @SIGMOD, 2024. (best short paper award 🥇)

Exploring the Numerical Reasoning Capabilities of Language Models: A Comprehensive Analysis on Tabular Data.
Mubashara Akhtar*, Abhilash Shankarampeta*, Vivek Gupta, Arpit Patil, Oana Cocarascu, Elena Simperl.
Findings of Annual Conference on Empirical Methods in Natural Language Processing (EMNLP), 2023.

Honours & Awards

Awarded the AI Center fellowship from ETH Zurich (~3% acceptance rate), 2024
Best short paper award 🥇 at DEEM 2024 for our paper introducing Croissant
King’s College London NMES Research Impact Award for my PhD research on multimodal fact-checking (across the Faculty of Natural, Mathematical and Engineering sciences), 2023
King’s College London NMES Faculty studentship, 2020-2024
European Forum Alpbach scholarship, 2020-2021
Selected for US Embassy Austria's Austria to Austin program, 2019
Ban Ki-moon Center mentoring partnership (mentored by Vice Rector of TU Wien, Anna Steiger), 2019
Research scholarship from TU Wien for my master's thesis project, 2017
Summer school and conference grant from TU Wien for ACM RecSys, 2017
Merit-based scholarship from TU Wien, 2012-2016
Top 1% with 1.0 overall score at Austrian high school final exam (summa cum laude)

News

Jan 2026: Two papers (LEXam and Bottleneck is Multimodal Reasoning) accepted at ICLR 2026.
Jan 2026: I'm moderating a roundtable on AI & Data at the AI House during the World Economic Forum in Davos.
Sept 2025: I'm co-organising the workshops FEVER at EACL 2026 and EvalEval at ACL 2026.
July 2025: I am attending ACL in Vienna to present TANQ (accepted at TACL) and co-host the FEVER workshop.
June 2025: Ev2R: Evidence Retrieval in Automated Fact-Checking was accepted at the TACL journal (updated preprint out soon).
Feb 2025: I was awarded the ETH AI Center fellowship (~3% acceptance rate) and started as a postdoc at ETH Zurich.
Jan 2025: My Google internship work TANQ: An open domain dataset of table answered questions was accepted at the journal TACL.
Dec 2024: I defended my thesis 🎉 & attended NeurIPS in Vancouver to present the Croissant metadata paper.
Sept 2024: Our paper on Croissant was accepted as a 🌟 spotlight paper at NeurIPS 2024 (top ~3% of submissions).
Aug 2024: Just submitted my PhD thesis! I'll be attending ACL 2024 in Bangkok to present ChartCheck.
June 2024: Our paper introducing the Croissant format for ML dataset documentation won the Best Short Paper Award at DEEM. 🎉
May 2024: Our paper on explainable fact-checking against chart images accepted at ACL (Findings).
Feb 2024: Croissant has been released! Check out the related publication, MLCommons announcement, and Google blog post.
Oct 2023: First-author papers on multimodal fact-checking and numerical reasoning with LLMs accepted at EMNLP 2023.
July 2023: I started an internship at Google Deepmind. I'll be in Zurich until November.
May 2023: Received a Research Impact Award at King's for my research on multimodal fact-checking.
May 2023: Attending EACL in Dubrovnik to present our paper on Chart Fact Verification and organising FEVER.
Jan 2023: Began a research visit at the University of Cambridge , working with Prof. Andreas Vlachos.
Jan 2023: Our paper Reading and Reasoning over Chart Images for Evidence-based Automated Fact-Checking was accepted to EACL.
May 2022: Attending ACL in Dublin and wrote a blog on Multimodality at ACL 2022.
April 2022: Our paper PubHealthTab: a Public Health Table-based Dataset for Evidence-based Fact Checking was accepted to NAACL.
Nov 2021: Talked at the FrauenFunk! podcast about my research, women in STEM, and doing a PhD in the UK. (German).
July 2021: I am a teaching assistant at the Oxford ML Summer School.
Dec 2020: Organising a WiML workshop at NeurIPS 2020.