Capital One at ACL 2026

Advancing AI and NLP Frontiers at ACL 2026

As language models grow more deeply integrated into technology ecosystems, pioneering robust, efficient, and reliable Natural Language Processing (NLP) techniques becomes paramount. Capital One continues to invest in state-of-the-art AI/ML science through deep multi-sector collaboration and peer-reviewed research. At the upcoming Annual Meeting of the Association for Computational Linguistics (ACL 2026), Capital One researchers and academic partners will showcase novel findings stretching from LLM security to multilingual capabilities.

Through the Science & Academic Partnerships program, Capital One bridges industry needs with academic expertise, funding critical university research and engineering solutions that make technology safer and more powerful. Our accepted publications at ACL 2026 demonstrate this thriving flywheel of talent and collaborative innovation across multiple research categories.

Main Conference Research

Adaptive Instruction Composition for Automated LLM Red Teaming
Capital One Authors: Jesse Zymet, Swapnil Shinde, Sahil Wadhwa, Andy Luo

Overview: Standard red teaming approaches often struggle with a limited range of jailbreak strategies or rely on ineffective, randomized crowd-sourced tactics. This paper introduces a novel framework—Adaptive Instruction Composition—that utilizes reinforcement learning and a neural contextual bandit to tailor attack compositions dynamically, balancing diversity and effectiveness to proactively uncover target model vulnerabilities.

Routing with Generated Data: Annotation-Free LLM Skill Estimation and Expert Selection
Capital One Authors: Genta Winata, Sambit Sahu, Supriyo Chakraborty, Shixiong Zhang

Overview: Emerging from our gifted research collaboration with University of North Carolina, Chapel Hill, this paper introduces Routing with Generated Data (RGD). It proposes CASCAL, a query-only router that accurately estimates model correctness via consensus voting and hierarchical clustering, outperforming traditional baselines even when trained on weaker generator datasets.

CommonLID: Re-evaluating State-of-the-Art Language Identification Performance on Web Data
Capital One Authors: Genta Winata

Overview: Addressing the gaps in multilingual corpora curation, this research introduces CommonLID—a community-driven, human-annotated benchmark covering 109 languages. The study reveals that existing evaluations often overestimate web-domain language identification accuracy, providing a vital tool for developing representative data.

Macaron: Controlled, Human-Written Benchmark for Multilingual and Multicultural Reasoning via Template-Filling 
Capital One Authors: Genta Winata

Overview: This paper introduces Macaron, a template-first benchmark factorizing reasoning type and cultural context across 20 languages and 10 scripts. Evaluating 21 multilingual LLMs, the benchmark highlights performance drops in open-weight models when working with local languages, identifying crucial spaces for future model training.

Workshop Track Research

Temporal Tokenization Strategies for Event Sequence Modeling with Large Language Models
Capital One Authors: Zefang Liu, Nam Nguyen, Shixiong Zhang
Track: GEM and SURGeLLM Workshops

Overview: Effectively representing continuous time is a major hurdle in event sequence modeling. A systematic empirical study is presented comparing encoding strategies (like calendar tokens and adaptive residual scalar quantization) across diverse statistical distributions, highlighting how critical tokenizer alignment is to prediction performance.

EconWebArena: Benchmarking Autonomous Agents on Economic Tasks in Realistic Web Environments
Capital One Authors: Zefang Liu
Track: GEM and SURGeLLM Workshops

Overview: A novel benchmark comprising 360 curated tasks from 82 authoritative websites is introduced to evaluate autonomous agents on complex economic tasks within realistic web environments. Through the evaluation of diverse multimodal LLMs, substantial performance gaps are revealed, highlighting persistent challenges in visual grounding, navigation, and multimodal understanding.

Connect with Capital One at ACL 2026

If you’re attending ACL, we invite you to visit our booth to engage with our researchers and authors.

• Visit our booth: D2
• Explore our research: Dive deep into our latest advancements in AI and machine learning.
• Discover career opportunities: Learn about exciting applied research career paths at Capital One for researchers and engineers passionate about AI and join our world-class team.
• Learn about our student and grad internships: Put your knowledge and skills to work in our 10-week to two-year graduate programs innovating new products and creatively solving the problems that impact our customers and our business.
• Engage with our team: Meet our researchers and AI experts, explore how we’re shaping financial services with patented AI and discuss what’s next for AI in finance.

Disclaimers & Disclosures

DISCLOSURE STATEMENT: © 2026 Capital One. Opinions are those of the individual author and are not necessarily those of Capital One. Unless noted otherwise, Capital One is not affiliated with, nor endorsed by, any third parties mentioned and is not responsible for the content or privacy policies of any linked third-party sites. Any trademarks and other intellectual property used or displayed are property of their respective owners.