# TWEAC: Transformer with Extendable QA Agent Classifiers

arXiv: https://arxiv.org/abs/2104.07081 · HTML: https://mlsanity.com/paper/2104.07081/

## Key facts

- Cited by 37 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 51 references that are in the map.
- First posted on arXiv 14 Apr 2021.

## Builds on (most cited first, 27 of 27)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [SQuAD: 100,000+ Questions for Machine Comprehension of Text](https://mlsanity.com/paper/1606.05250/) (2016)
- [HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering](https://mlsanity.com/paper/1809.09600/) (2018)
- [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://mlsanity.com/paper/1909.11942/) (2019)
- [BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questions](https://mlsanity.com/paper/1905.10044/) (2019)
- [DROP: A Reading Comprehension Benchmark Requiring Discrete Reasoning Over Paragraphs](https://mlsanity.com/paper/1903.00161/) (2019)
- [Reading Wikipedia to Answer Open-Domain Questions](https://mlsanity.com/paper/1704.00051/) (2017)
- [Snips Voice Platform: an embedded Spoken Language Understanding system for private-by-design voice interfaces](https://mlsanity.com/paper/1805.10190/) (2018)
- [Multilingual Universal Sentence Encoder for Semantic Retrieval](https://mlsanity.com/paper/1907.04307/) (2019)
- [PullNet: Open Domain Question Answering with Iterative Retrieval on Knowledge Bases and Text](https://mlsanity.com/paper/1904.09537/) (2019)
- [Open Domain Question Answering Using Early Fusion of Knowledge Bases and Text](https://mlsanity.com/paper/1809.00782/) (2018)
- [End-to-End Open-Domain Question Answering with BERTserini](https://mlsanity.com/paper/1902.01718/) (2019)
- [ParlAI: A Dialog Research Software Platform](https://mlsanity.com/paper/1705.06476/) (2017)
- [Attention-Based Recurrent Neural Network Models for Joint Intent Detection and Slot Filling](https://mlsanity.com/paper/1609.01454/) (2016)
- [A Stack-Propagation Framework with Token-Level Intent Detection for Spoken Language Understanding](https://mlsanity.com/paper/1909.02188/) (2019)
- [A Richly Annotated Corpus for Different Tasks in Automated Fact-Checking](https://mlsanity.com/paper/1911.01214/) (2019)
- [MultiQA: An Empirical Investigation of Generalization and Transfer in Reading Comprehension](https://mlsanity.com/paper/1905.13453/) (2019)
- [KBQA: Learning Question Answering over QA Corpora and Knowledge Bases](https://mlsanity.com/paper/1903.02419/) (2019)
- [MultiReQA: A Cross-Domain Evaluation for Retrieval Question Answering Models](https://mlsanity.com/paper/2005.02507/) (2020)
- [MultiCQA: Zero-Shot Transfer of Self-Supervised Text Matching Models on a Massive Scale](https://mlsanity.com/paper/2010.00980/) (2020)
- [WikiPassageQA: A Benchmark Collection for Research on Non-factoid Answer Passage Retrieval](https://mlsanity.com/paper/1805.03797/) (2018)
- [Learning to Rank Question Answer Pairs with Holographic Dual LSTM Architecture](https://mlsanity.com/paper/1707.06372/) (2017)
- [PaRe: A Paper-Reviewer Matching Approach Using a Common Topic Space](https://mlsanity.com/paper/1909.11258/) (2019)
- [SemEval-2017 Task 3: Community Question Answering](https://mlsanity.com/paper/1912.00730/) (2019)
- [Continuous Learning for Large-scale Personalized Domain Classification](https://mlsanity.com/paper/1905.00921/) (2019)
- [Neural Duplicate Question Detection without Labeled Training Data](https://mlsanity.com/paper/1911.05594/) (2019)

## Built on by (most cited first, 30 of 37)

- [MTEB: Massive Text Embedding Benchmark](https://mlsanity.com/paper/2210.07316/) (2022)
- [NV-Embed: Improved Techniques for Training LLMs as Generalist Embedding Models](https://mlsanity.com/paper/2405.17428/) (2024)
- [Large Language Model as Attributed Training Data Generator: A Tale of Diversity and Bias](https://mlsanity.com/paper/2306.15895/) (2023)
- [ClusterLLM: Large Language Models as a Guide for Text Clustering](https://mlsanity.com/paper/2305.14871/) (2023)
- [Llama-Embed-Nemotron-8B: A Universal Text Embedding Model for Multilingual and Cross-Lingual Tasks](https://mlsanity.com/paper/2511.07025/) (2025)
- [KaLM-Embedding-V2: Superior Training Techniques and Data Inspire A Versatile Embedding Model](https://mlsanity.com/paper/2506.20923/) (2025)
- [Generative AI for Synthetic Data Generation: Methods, Challenges and the Future](https://mlsanity.com/paper/2403.04190/) (2024)
- [Beyond Matryoshka: Revisiting Sparse Coding for Adaptive Representation](https://mlsanity.com/paper/2503.01776/) (2025)
- [MetaQA: Combining Expert Agents for Multi-Skill Question Answering](https://mlsanity.com/paper/2112.01922/) (2021)
- [SynthesizRR: Generating Diverse Datasets with Retrieval Augmentation](https://mlsanity.com/paper/2405.10040/) (2024)
- [F2LLM-v2: Inclusive, Performant, and Efficient Embeddings for a Multilingual World](https://mlsanity.com/paper/2603.19223/) (2026)
- [Redundancy, Isotropy, and Intrinsic Dimensionality of Prompt-based Text Embeddings](https://mlsanity.com/paper/2506.01435/) (2025)
- [Length-Induced Embedding Collapse in PLM-based Models](https://mlsanity.com/paper/2410.24200/) (2024)
- [F2LLM Technical Report: Matching SOTA Embedding Performance with 6 Million Open-Source Data](https://mlsanity.com/paper/2510.02294/) (2025)
- [HUME: Measuring the Human-Model Performance Gap in Text Embedding Tasks](https://mlsanity.com/paper/2510.10062/) (2025)
- [UKP-SQUARE: An Online Platform for Question Answering Research](https://mlsanity.com/paper/2203.13693/) (2022)
- [CSRv2: Unlocking Ultra-Sparse Embeddings](https://mlsanity.com/paper/2602.05735/) (2026)
- [Bilingual BSARD: Extending Statutory Article Retrieval to Dutch](https://mlsanity.com/paper/2412.07462/) (2024)
- [Attributes as Textual Genes: Leveraging LLMs as Genetic Algorithm Simulators for Conditional Synthetic Data Generation](https://mlsanity.com/paper/2509.02040/) (2025)
- [Quantifying Positional Biases in Text Embedding Models](https://mlsanity.com/paper/2412.15241/) (2024)
- [Conan-Embedding-v2: Training an LLM from Scratch for Text Embeddings](https://mlsanity.com/paper/2509.12892/) (2025)
- [LGAI-EMBEDDING-Preview Technical Report](https://mlsanity.com/paper/2506.07438/) (2025)
- [On The Role of Pretrained Language Models in General-Purpose Text Embeddings: A Survey](https://mlsanity.com/paper/2507.20783/) (2025)
- [One prompt is not enough: Instruction Sensitivity Undermines Embedding Model Evaluation](https://mlsanity.com/paper/2605.22544/) (2026)
- [UKP-SQuARE v2: Explainability and Adversarial Attacks for Trustworthy QA](https://mlsanity.com/paper/2208.09316/) (2022)
- [Non-contrastive sentence representations via self-supervision](https://mlsanity.com/paper/2310.17690/) (2023)
- [Continuous Risk Prediction](https://mlsanity.com/paper/2410.09449/) (2024)
- [UKP-SQuARE v3: A Platform for Multi-Agent QA Research](https://mlsanity.com/paper/2303.18120/) (2023)
- [Learning Concepts, Not Tokens: Self-Supervised Semantic Alignment for Language Models](https://mlsanity.com/paper/2603.29123/) (2026)
- [Vector Linking via Cross-Model Local Isometric Consistency](https://mlsanity.com/paper/2605.31100/) (2026)

## Abstract (arXiv)

Question answering systems should help users to access knowledge on a broad range of topics and to answer a wide array of different questions. Most systems fall short of this expectation as they are only specialized in one particular setting, e.g., answering factual questions with Wikipedia data. To overcome this limitation, we propose composing multiple QA agents within a meta-QA system. We argue that there exist a wide range of specialized QA agents in literature. Thus, we address the central research question of how to effectively and efficiently identify suitable QA agents for any given question. We study both supervised and unsupervised approaches to address this challenge, showing that TWEAC -- Transformer with Extendable Agent Classifiers -- achieves the best performance overall with 94% accuracy. We provide extensive insights on the scalability of TWEAC, demonstrating that it scales robustly to over 100 QA agents with each providing just 1000 examples of questions they can answer. Our code and data is available: https://github.com/UKPLab/TWEAC-qa-agent-selection