• Home
• Publications
• Resume

Cong Duy Vu Hoang, PhD

Research & Development — on AI systems by day, on recipes and on myself the rest of the time.

Senior Manager, Applied Science & AI
Oracle Corporation
Melbourne, VIC, Australia

Email
Google Scholar
GitHub
LinkedIn

About

I am a Senior Manager of Applied Science & AI at Oracle Corporation, Melbourne, Australia. My research and engineering work spans Natural Language Processing (NLP), Generative AI, Large Language Models (LLMs), Agentic AI, NL2SQL, NL2Code, Conversational AI, Analytics AI, and Healthcare AI. I lead applied science teams building production-grade AI systems across Oracle's enterprise cloud portfolio — Oracle Analytics Cloud (OAC) and Oracle Health & AI (OHAI).

I obtained my PhD in Engineering (NLP & Deep Learning) from the University of Melbourne in 2019, under the joint supervision of Prof. Trevor Cohn and Prof. Reza Haffari. Prior to Oracle, I was an AI Scientist at Speak.AI (a spin-off of Voicebox Technologies, subsequently acquired by Oracle), a research intern at NAVER LABS Europe (formerly Xerox Research Centre Europe), a visiting scholar at Carnegie Mellon University, Language Technologies Institute, and a Senior Research Engineer at HLT, I²R, A*STAR, Singapore. Before that, I studied at National University of Singapore (NUS) (MSc) and was a teaching & research assistant at the University of Science, Vietnam National University HCMC.

I hold 38 granted U.S. patents and 45 pending, covering NL2SQL, NL2Code, LLM training, NER, and agentic systems. My research has received 1,380+ citations (h-index 14, i10-index 19) on Google Scholar, with publications at ACL, EMNLP, NAACL, and COLING.

Recent Highlights

*** Released two new curated repos on GitHub: awesome-reliable-text-to-sql (papers, systems, and datasets on Text-to-SQL ambiguity, clarification, unanswerability, and abstention) and awesome-ai-co-scientist (comparisons, reproducibility reports, and evaluation recipes for AI co-scientist systems).

*** Our paper CLARITY: A Framework and Benchmark for Conversational Language Ambiguity and Unanswerability in Interactive NL2SQL Systems has been accepted at ACL 2026 (Industry Track) --- code released.

*** Two papers accepted at NAACL 2025: Distill-C: Enhanced NL2SQL via Distilled Customization with LLMs (Industry Track) and Mastering the Craft of Data Synthesis for CodeLLMs (Long Paper).

*** Our paper SQLong: Enhanced NL2SQL for Longer Contexts with LLMs has been accepted at the Table Representation Learning Workshop at ACL 2025.

*** I have transitioned to Senior Manager, Applied Science & AI at Oracle Corporation (Oct 2024), leading applied science teams across OAC and OHAI.

*** Our Oracle IP portfolio has grown to 38 granted U.S. patents and 45 pending applications spanning NL2SQL, NL2Code, LLM training, NER, and dialog systems.

Selected Patents (2022–2026)

See full patent portfolio (38 granted • 45 pending).

Selected Publications

CLARITY: A Framework and Benchmark for Conversational Language Ambiguity and Unanswerability in Interactive NL2SQL Systems
Tabinda Sarwar, Farhad Moghimifar, Cong Duy Vu Hoang, Xiaoxiao Ma, Shawn Chang Xu, Fahimeh Saleh, Poorya Zaremoodi, Avirup Sil, Katrin Kirchhoff. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026) (Industry Track), 2026. NL2SQL
Abstract

×

CLARITY: A Framework and Benchmark for Conversational Language Ambiguity and Unanswerability in Interactive NL2SQL Systems

We introduce CLARITY, a framework and benchmark for evaluating conversational language ambiguity and unanswerability in interactive NL2SQL systems. Real-world NL2SQL deployments must handle ambiguous, underspecified, or unanswerable user requests in a multi-turn dialogue setting. CLARITY provides a systematic taxonomy of ambiguity types and unanswerability conditions, along with a benchmark dataset for evaluating how well NL2SQL systems identify, clarify, and resolve such inputs.

Code

Distill-C: Enhanced NL2SQL via Distilled Customization with LLMs
Cong Duy Vu Hoang, Gioacchino Tangari, Clemence Lanfranchi, Dalu Guo, Paul Cayet, Steve Siu, Don Dharmasiri, Yuan-Fang Li, Long Duong, Damien Hilloulin, Rhicheek Patra, Sungpack Hong, Hassan Chafi. In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics (NAACL 2025) (Industry Track), 2025. NL2SQL
Abstract

×

Distill-C: Enhanced NL2SQL via Distilled Customization with LLMs

We present Distill-C, a framework for enhancing NL2SQL systems via distilled customization with large language models. By combining knowledge distillation with domain-specific customization, Distill-C achieves strong NL2SQL performance while remaining practical for enterprise deployment across Oracle SQL and PL/SQL environments.

Mastering the Craft of Data Synthesis for CodeLLMs
Meng Chen, Philip Arthur, Qianyu Feng, Cong Duy Vu Hoang, Yu-Heng Hong, Mahdi Kazemi Moghaddam, Omid Nezami, Duc Thien Nguyen, Gioacchino Tangari, Duy Vu, Thanh Vu, Mark Johnson, Krishnaram Kenthapadi, Don Dharmasiri, Long Duong, Yuan-Fang Li. In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics (NAACL 2025) (Long Paper), 2025. NL2Code
Abstract

×

Mastering the Craft of Data Synthesis for CodeLLMs

We present a systematic study on data synthesis for training Code LLMs. We identify key principles for crafting high-quality synthetic training data and demonstrate their effectiveness across code generation benchmarks. Our findings provide practical guidance for building production-grade Code LLMs with limited human-annotated data.

On the Use of Prior and External Knowledge in Neural Sequence Models
PhD Thesis, The University of Melbourne, VIC, Australia. 2019. PhD Thesis
Abstract

×

On the Use of Prior and External Knowledge in Neural Sequence Models

Neural sequence models have recently achieved great success across various natural language processing tasks. In practice, neural sequence models require massive amount of annotated training data to reach their desirable performance; however, there will not always be available data across languages, domains or tasks at hand. Prior and external knowledge provides additional contextual information, potentially improving the modelling performance as well as compensating the lack of large training data, particular in low-resourced situations. In this thesis, we investigate the usefulness of utilising prior and external knowledge for improving neural sequence models. We propose the use of various kinds of prior and external knowledge and present different approaches for integrating them into both training and inference phases of neural sequence models. The followings are main contributions of this thesis which are summarised in two major parts. We present the first part of this thesis which is on Training and Modelling for neural sequence models. In this part, we investigate different situations (particularly in low resource settings) in which prior and external knowledge, such as side information, linguistic factors, monolingual data, is shown to have great benefits for improving performance of neural sequence models. In addition, we introduce a new means for incorporating prior and external knowledge based on the moment matching framework. This framework serves its purpose for exploiting prior and external knowledge as global features of generated sequences in neural sequence models in order to improve the overall quality of the desired output sequence. The second part is about Decoding of neural sequence models in which we propose a novel decoding framework with relaxed continuous optimisation in order to address one of the drawbacks of existing approximate decoding methods, namely the limited ability to incorporate global factors due to intractable search. We hope that this PhD thesis, constituted by two above major parts, will shed light on the use of prior and external knowledge in neural sequence models, both in their training and decoding phases.

An Adaptable Task-oriented Dialog System for Stand-alone Embedded Devices
Long Duong, Cong Duy Vu Hoang, Tuyen Quang Pham, Yu-Heng Hong, Vladislavs Dovgalecs, Guy Bashkansky, Jason Black, Andrew Bleeker, Serge Le Huitouze, Mark Johnson. In Proceedings of Annual Meeting of the Association for Computational Linguistics (ACL-19) (System Demonstrations), 2019. Dialogue
Abstract

×

An Adaptable Task-oriented Dialog System for Stand-alone Embedded Devices

This paper describes a spoken-language end-to-end task-oriented dialogue system for small embedded devices such as home appliances. While the current system implements a smart alarm clock with advanced calendar scheduling functionality, the system is designed to make it easy to port to other application domains (e.g., the dialogue component factors out domain-specific execution from domain-general actions such as requesting and updating slot values). The system does not require internet connectivity because all components, including speech recognition, natural language understanding, dialogue management, execution and text-to-speech, run locally on the embedded device (our demo uses a Raspberry Pi). This simplifies deployment, minimizes server costs and most importantly, eliminates user privacy risks. The demo video in alarm domain is here youtu.be/N3IBMGocvHU.

Video

Moment Matching Training for Neural Machine Translation - A Preliminary Study
Cong Duy Vu Hoang, Ioan Calapodescu, Marc Dymetman. In arXiv preprint, 2018. Machine Translation
Abstract

×

Moment Matching Training for Neural Machine Translation - A Preliminary Study

In previous works, neural sequence models have been shown to improve significantly if external prior knowledge can be provided, for instance by allowing the model to access the embeddings of explicit features during both training and inference. In this work, we propose a different point of view on how to incorporate prior knowledge in a principled way, using a moment matching framework. In this approach, the standard local cross-entropy training of the sequential model is combined with a moment matching training mode that encourages the equality of the expectations of certain predefined features between the model distribution and the empirical distribution. In particular, we show how to derive unbiased estimates of some stochastic gradients that are central to the training, and compare our framework with a formally related one: policy gradient training in reinforcement learning, pointing out some important differences in terms of the kinds of prior assumptions in both approaches. Our initial results are promising, showing the effectiveness of our proposed framework.

Improved Neural Machine Translation using Side Information
Cong Duy Vu Hoang, Gholamreza Haffari and Trevor Cohn. In Proceedings of The 16th Annual Workshop of The Australasian Language Technology Association (ALTA'18) (long, oral) (best paper award), 2018. Machine Translation
Abstract

×

Improved Neural Machine Translation using Side Information

In this work, we investigate whether side information is helpful in the context of neural machine translation (NMT). We study various kinds of side information, including topical information and personal traits, and then propose different ways of incorporating these information sources into existing NMT models. Our experimental results show the benefits of side in- formation in improving the NMT models.

Code

Iterative Back-Translation for Neural Machine Translation
Cong Duy Vu Hoang, Philipp Koehn, Gholamreza Haffari and Trevor Cohn. In Proceedings of The 2nd Workshop on Neural Machine Translation and Generation associated with ACL 2018 (long, poster), 2018. Machine Translation
Abstract

×

Iterative Back-Translation for Neural Machine Translation

We present iterative back-translation, a method for generating increasingly better synthetic parallel data from monolingual data to train neural machine translation systems. We demonstrate improvements in neural machine translation quality in both high and low resourced scenarios, including the best reported BLEU scores for the WMT 2017 German$English tasks.

Towards Decoding as Continuous Optimization in Neural Machine Translation
Cong Duy Vu Hoang, Gholamreza Haffari and Trevor Cohn. In Proceedings of Conference on Empirical Methods in Natural Language Processing (EMNLP'17) (long, oral), 2017. Machine Translation
Abstract

×

Towards Decoding as Continuous Optimization in Neural Machine Translation

We propose a novel decoding approach for neural machine translation (NMT) based on continuous optimisation. We reformulate decoding, a discrete optimization problem, into a continuous problem, such that optimization can make use of efficient gradient-based techniques. Our powerful decoding framework allows for more accurate decoding for standard neural machine translation models, as well as enabling decoding in intractable models such as intersection of several different NMT models. Our empirical results show that our decoding framework is effective, and can leads to substantial improvements in translations, especially in situations where greedy search and beam search are not feasible. Finally, we show how the technique is highly competitive with, and complementary to, reranking.

Code

Improving Neural Translation Models with Linguistic Factors
Cong Duy Vu Hoang, Gholamreza Haffari and Trevor Cohn. In Proceedings of The 14th Annual Workshop of The Australasian Language Technology Association (ALTA'16) (long, oral) (best paper award), 2016. Machine Translation
Abstract

×

Improving Neural Translation Models with Linguistic Factors

This paper presents an extension of neural machine translation (NMT) model to incorporate additional word-level linguistic factors. Adding such linguistic factors may be of great benefits to learning of NMT models, potentially reducing language ambiguity or alleviating data sparseness problem (Koehn and Hoang, 2007). We explore different linguistic annotations at the word level, including: lemmatization, word clusters, Part-of-Speech tags, and labeled dependency relations. We then propose different neural attention architectures to integrate these additional factors into the NMT framework. Evaluating on translating between English and German in two directions with a low resource setting in the domain of TED talks, we obtain promising results in terms of both perplexity reductions and improved BLEU scores over baseline methods.

Incorporating Structural Alignment Biases into an Attentional Neural Translation Model
Trevor Cohn, Cong Duy Vu Hoang, Ekaterina Vylomova, Kaisheng Yao, Chris Dyer and Gholamreza Haffari. In Proceedings of The 15th Annual Conference of the North American Chapter of the Association for Computational Linguistics - Human Language Technologies (NAACL-HLT'16) (long), 2016. Machine Translation
Abstract

×

Incorporating Structural Alignment Biases into an Attentional Neural Translation Model

Neural encoder-decoder models of machine translation have achieved impressive results, rivalling traditional translation models. However their modelling formulation is overly simplistic, and omits several key inductive biases built into traditional models. In this paper we extend the attentional neural translation model to include structural biases from word based alignment models, including positional bias, Markov conditioning, fertility and agreement over translation directions. We show improvements over a baseline attentional model and standard phrase-based model over several language pairs, evaluating on difficult languages in a low resource setting.

Code

Incorporating Side Information into Recurrent Neural Network Language Models
Cong Duy Vu Hoang, Gholamreza Haffari and Trevor Cohn. In Proceedings of The 15th Annual Conference of the North American Chapter of the Association for Computational Linguistics - Human Language Technologies (NAACL-HLT'16) (short), 2016. Language Modeling
Abstract

×

Incorporating Side Information into Recurrent Neural Network Language Models

Recurrent neural network language models (RNNLM) have recently demonstrated vast potential in modelling long-term dependencies for NLP problems, ranging from speech recognition to machine translation. In this work, we propose methods for conditioning RNNLMs on external side information, e.g., metadata such as keywords or document title. Our experiments show consistent improvements of RNNLMs using side information over the baselines for two different datasets and genres in two languages. Interestingly, we found that side information in a foreign language can be highly beneficial in modelling texts in another language, serving as a form of cross-lingual language modelling.

Code