Tung Nguyen

Tung Nguyen, Johannes Brandstetter, Ashish Kapoor et al.

Most state-of-the-art approaches for weather and climate modeling are based on physics-informed numerical models of the atmosphere. These approaches aim to model the non-linear dynamics and complex interactions between multiple variables, which are challenging to approximate. Additionally, many such numerical models are computationally intensive, especially when modeling the atmospheric phenomenon at a fine-grained spatial and temporal resolution. Recent data-driven approaches based on machine learning instead aim to directly solve a downstream forecasting or projection task by learning a data-driven functional mapping using deep neural networks. However, these networks are trained using curated and homogeneous climate datasets for specific spatiotemporal tasks, and thus lack the generality of numerical models. We develop and demonstrate ClimaX, a flexible and generalizable deep learning model for weather and climate science that can be trained using heterogeneous datasets spanning different variables, spatio-temporal coverage, and physical groundings. ClimaX extends the Transformer architecture with novel encoding and aggregation blocks that allow effective use of available compute while maintaining general utility. ClimaX is pre-trained with a self-supervised learning objective on climate datasets derived from CMIP6. The pre-trained ClimaX can then be fine-tuned to address a breadth of climate and weather tasks, including those that involve atmospheric variables and spatio-temporal scales unseen during pretraining. Compared to existing data-driven baselines, we show that this generality in ClimaX results in superior performance on benchmarks for weather forecasting and climate projections, even when pretrained at lower resolutions and compute budgets. The source code is available at https://github.com/microsoft/ClimaX.

Tung Nguyen, Jason Jewik, Hritik Bansal et al.

Modeling weather and climate is an essential endeavor to understand the near- and long-term impacts of climate change, as well as inform technology and policymaking for adaptation and mitigation efforts. In recent years, there has been a surging interest in applying data-driven methods based on machine learning for solving core problems such as weather forecasting and climate downscaling. Despite promising results, much of this progress has been impaired due to the lack of large-scale, open-source efforts for reproducibility, resulting in the use of inconsistent or underspecified datasets, training setups, and evaluations by both domain scientists and artificial intelligence researchers. We introduce ClimateLearn, an open-source PyTorch library that vastly simplifies the training and evaluation of machine learning models for data-driven climate science. ClimateLearn consists of holistic pipelines for dataset processing (e.g., ERA5, CMIP6, PRISM), implementation of state-of-the-art deep learning models (e.g., Transformers, ResNets), and quantitative and qualitative evaluation for standard weather and climate modeling tasks. We supplement these functionalities with extensive documentation, contribution guides, and quickstart tutorials to expand access and promote community growth. We have also performed comprehensive forecasting and downscaling experiments to showcase the capabilities and key features of our library. To our knowledge, ClimateLearn is the first large-scale, open-source effort for bridging research in weather and climate modeling with modern machine learning systems. Our library is available publicly at https://github.com/aditya-grover/climate-learn.

Tung Nguyen, Aditya Grover

Neural Processes (NPs) are a popular class of approaches for meta-learning. Similar to Gaussian Processes (GPs), NPs define distributions over functions and can estimate uncertainty in their predictions. However, unlike GPs, NPs and their variants suffer from underfitting and often have intractable likelihoods, which limit their applications in sequential decision making. We propose Transformer Neural Processes (TNPs), a new member of the NP family that casts uncertainty-aware meta learning as a sequence modeling problem. We learn TNPs via an autoregressive likelihood-based objective and instantiate it with a novel transformer-based architecture. The model architecture respects the inductive biases inherent to the problem structure, such as invariance to the observed data points and equivariance to the unobserved points. We further investigate knobs within the TNP framework that tradeoff expressivity of the decoding distribution with extra computation. Empirically, we show that TNPs achieve state-of-the-art performance on various benchmark problems, outperforming all previous NP variants on meta regression, image completion, contextual multi-armed bandits, and Bayesian optimization.

Tung Nguyen, Sudhanshu Agrawal, Aditya Grover

Experimental design is a fundamental problem in many science and engineering fields. In this problem, sample efficiency is crucial due to the time, money, and safety costs of real-world design evaluations. Existing approaches either rely on active data collection or access to large, labeled datasets of past experiments, making them impractical in many real-world scenarios. In this work, we address the more challenging yet realistic setting of few-shot experimental design, where only a few labeled data points of input designs and their corresponding values are available. We approach this problem as a conditional generation task, where a model conditions on a few labeled examples and the desired output to generate an optimal input design. To this end, we introduce Experiment Pretrained Transformers (ExPT), a foundation model for few-shot experimental design that employs a novel combination of synthetic pretraining with in-context learning. In ExPT, we only assume knowledge of a finite collection of unlabelled data points from the input domain and pretrain a transformer neural network to optimize diverse synthetic functions defined over this domain. Unsupervised pretraining allows ExPT to adapt to any design task at test time in an in-context fashion by conditioning on a few labeled data points from the target task and generating the candidate optima. We evaluate ExPT on few-shot experimental design in challenging domains and demonstrate its superior generality and performance compared to existing methods. The source code is available at https://github.com/tung-nd/ExPT.git.

Tung Nguyen, Qinqing Zheng, Aditya Grover

Behavioral cloning (BC) provides a straightforward solution to offline RL by mimicking offline trajectories via supervised learning. Recent advances (Chen et al., 2021; Janner et al., 2021; Emmons et al., 2021) have shown that by conditioning on desired future returns, BC can perform competitively to their value-based counterparts, while enjoying much more simplicity and training stability. While promising, we show that these methods can be unreliable, as their performance may degrade significantly when conditioned on high, out-of-distribution (ood) returns. This is crucial in practice, as we often expect the policy to perform better than the offline dataset by conditioning on an ood value. We show that this unreliability arises from both the suboptimality of training data and model architectures. We propose ConserWeightive Behavioral Cloning (CWBC), a simple and effective method for improving the reliability of conditional BC with two key components: trajectory weighting and conservative regularization. Trajectory weighting upweights the high-return trajectories to reduce the train-test gap for BC methods, while conservative regularizer encourages the policy to stay close to the data distribution for ood conditioning. We study CWBC in the context of RvS (Emmons et al., 2021) and Decision Transformers (Chen et al., 2021), and show that CWBC significantly boosts their performance on various benchmarks.

Tung Nguyen, Jeffrey Uhlmann

We introduce a new consistency-based approach for defining and solving nonnegative/positive matrix and tensor completion problems. The novelty of the framework is that instead of artificially making the problem well-posed in the form of an application-arbitrary optimization problem, e.g., minimizing a bulk structural measure such as rank or norm, we show that a single property/constraint: preserving unit-scale consistency, guarantees the existence of both a solution and, under relatively weak support assumptions, uniqueness. The framework and solution algorithms also generalize directly to tensors of arbitrary dimensions while maintaining computational complexity that is linear in problem size for fixed dimension d. In the context of recommender system (RS) applications, we prove that two reasonable properties that should be expected to hold for any solution to the RS problem are sufficient to permit uniqueness guarantees to be established within our framework. This is remarkable because it obviates the need for heuristic-based statistical or AI methods despite what appear to be distinctly human/subjective variables at the heart of the problem. Key theoretical contributions include a general unit-consistent tensor-completion framework with proofs of its properties, e.g., consensus-order and fairness, and algorithms with optimal runtime and space complexities, e.g., O(1) term-completion with preprocessing complexity that is linear in the number of known terms of the matrix/tensor. From a practical perspective, the seamless ability of the framework to generalize to exploit high-dimensional structural relationships among key state variables, e.g., user and product attributes, offers a means for extracting significantly more information than is possible for alternative methods that cannot generalize beyond direct user-product relationships.

Tung Nguyen, Rohan Shah, Hritik Bansal et al.

Weather forecasting is a fundamental problem for anticipating and mitigating the impacts of climate change. Recently, data-driven approaches for weather forecasting based on deep learning have shown great promise, achieving accuracies that are competitive with operational systems. However, those methods often employ complex, customized architectures without sufficient ablation analysis, making it difficult to understand what truly contributes to their success. Here we introduce Stormer, a simple transformer model that achieves state-of-the-art performance on weather forecasting with minimal changes to the standard transformer backbone. We identify the key components of Stormer through careful empirical analyses, including weather-specific embedding, randomized dynamics forecast, and pressure-weighted loss. At the core of Stormer is a randomized forecasting objective that trains the model to forecast the weather dynamics over varying time intervals. During inference, this allows us to produce multiple forecasts for a target lead time and combine them to obtain better forecast accuracy. On WeatherBench 2, Stormer performs competitively at short to medium-range forecasts and outperforms current methods beyond 7 days, while requiring orders-of-magnitude less training data and compute. Additionally, we demonstrate Stormer's favorable scaling properties, showing consistent improvements in forecast accuracy with increases in model size and training tokens. Code and checkpoints are available at https://github.com/tung-nd/stormer.

Duy-Tung Pham, Thien Trang Nguyen Vu, Tung Nguyen et al.

Recent advances in neural topic models have concentrated on two primary directions: the integration of the inference network (encoder) with a pre-trained language model (PLM) and the modeling of the relationship between words and topics in the generative model (decoder). However, the use of large PLMs significantly increases inference costs, making them less practical for situations requiring low inference times. Furthermore, it is crucial to simultaneously model the relationships between topics and words as well as the interrelationships among topics themselves. In this work, we propose a novel framework called NeuroMax (Neural Topic Model with Maximizing Mutual Information with Pretrained Language Model and Group Topic Regularization) to address these challenges. NeuroMax maximizes the mutual information between the topic representation obtained from the encoder in neural topic models and the representation derived from the PLM. Additionally, NeuroMax employs optimal transport to learn the relationships between topics by analyzing how information is transported among them. Experimental results indicate that NeuroMax reduces inference time, generates more coherent topics and topic groups, and produces more representative document embeddings, thereby enhancing performance on downstream tasks.

Hung Viet Pham, Tung Thanh Nguyen

Traditional defect prediction approaches often use metrics that measure the complexity of the design or implementing code of a software system, such as the number of lines of code in a source file. In this paper, we explore a different approach based on content of source code. Our key assumption is that source code of a software system contains information about its technical aspects and those aspects might have different levels of defect-proneness. Thus, content-based features such as words, topics, data types, and package names extracted from a source code file could be used to predict its defects. We have performed an extensive empirical evaluation and found that: i) such content-based features have higher predictive power than code complexity metrics and ii) the use of feature selection, reduction, and combination further improves the prediction performance.

Tung Nguyen, Mona Bavarian

Here, we demonstrate how machine learning enables the prediction of comonomers reactivity ratios based on the molecular structure of monomers. We combined multi-task learning, multi-inputs, and Graph Attention Network to build a model capable of predicting reactivity ratios based on the monomers chemical structures.

Hritik Bansal, Daniel Israel, Siyan Zhao et al.

Recent advancements in mixed-modal generative have opened new avenues for developing unified biomedical assistants capable of analyzing biomedical images, answering complex questions about them, and generating multimodal patient reports. However, existing datasets face challenges such as small sizes, limited coverage of biomedical tasks and domains, and a reliance on narrow sources. To address these gaps, we present MedMax, a large-scale multimodal biomedical instruction-tuning dataset for mixed-modal foundation models. With 1.47 million instances, MedMax encompasses a diverse range of tasks, including interleaved image-text generation, biomedical image captioning and generation, visual chat, and report understanding. These tasks span knowledge across diverse biomedical domains, including radiology and histopathology, grounded in medical papers and YouTube videos. Subsequently, we fine-tune a mixed-modal foundation model on the MedMax dataset, achieving significant performance improvements: a 26% gain over the Chameleon model and an 18.3% improvement over GPT-4o across 12 downstream biomedical visual question-answering tasks. Finally, we introduce a unified evaluation suite for biomedical tasks to guide the development of mixed-modal biomedical AI assistants. The data, model, and code is available at https://mint-medmax.github.io/.

Tung Nguyen, Johane Takeuchi

The partially observable Markov decision process (POMDP) framework is a common approach for decision making under uncertainty. Recently, multiple studies have shown that by integrating relevant domain knowledge into POMDP belief estimation, we can improve the learned policy's performance. In this study, we propose a novel method for integrating the domain knowledge into probabilistic belief update in POMDP framework using Jeffrey's rule and normalization. We show that the domain knowledge can be utilized to reduce the data requirement and improve performance for POMDP policy learning with RL.

Phong Minh Vu, Tung Thanh Nguyen

Software security incidents occur everyday and thousands of software security reports are announced each month. Thus, it is difficult for software security researchers, engineers, and other stakeholders to follow software security topics of their interests in real-time. In this paper, we propose, SOSK, a novel tool for this problem. SOSK allows a user to import a collection of software security reports. It pre-processes and extracts the most important keywords from the textual description of the reports. Based on the similarity of embedding vectors of keywords, SOSK can expand and/or refine a keyword set from a much smaller set of user-provided keywords. Thus, SOSK allows users to define any topic of their interests and retrieve security reports relevant to that topic effectively. Our preliminary evaluation shows that SOSK can expand keywords and retrieve reports relevant to user requests.

Tien Phat Nguyen, Vu Minh Ngo, Tung Nguyen et al.

Cross-lingual topic modeling aims to uncover shared semantic themes across languages. Several methods have been proposed to address this problem, leveraging both traditional and neural approaches. While previous methods have achieved some improvements in topic diversity, they often struggle to ensure high topic coherence and consistent alignment across languages. We propose XTRA (Cross-Lingual Topic Modeling with Topic and Representation Alignments), a novel framework that unifies Bag-of-Words modeling with multilingual embeddings. XTRA introduces two core components: (1) representation alignment, aligning document-topic distributions via contrastive learning in a shared semantic space; and (2) topic alignment, projecting topic-word distributions into the same space to enforce crosslingual consistency. This dual mechanism enables XTRA to learn topics that are interpretable (coherent and diverse) and well-aligned across languages. Experiments on multilingual corpora confirm that XTRA significantly outperforms strong baselines in topic coherence, diversity, and alignment quality. Code and reproducible scripts are available at https: //github.com/tienphat140205/XTRA.

Zijian Ding, Tung Nguyen, Weikai Li et al.

Deep learning-based prediction models for High-Level Synthesis (HLS) of hardware designs often struggle to generalize. In this paper, we study how to close the generalizability gap of these models through pretraining on synthetic data and introduce Iceberg, a synthetic data augmentation approach that expands both large language model (LLM)-generated programs and weak labels of unseen design configurations. Our weak label generation method is integrated with an in-context model architecture, enabling meta-learning from actual and proximate labels. Iceberg improves the geometric mean modeling accuracy by $86.4\%$ when adapt to six real-world applications with few-shot examples and achieves a $2.47\times$ and a $1.12\times$ better offline DSE performance when adapting to two different test datasets. Our open-sourced code is here: https://github.com/UCLA-VAST/iceberg

Quang Duc Nguyen, Tung Nguyen, Duc Anh Nguyen et al.

Uncovering hidden topics from short texts is challenging for traditional and neural models due to data sparsity, which limits word co-occurrence patterns, and label sparsity, stemming from incomplete reconstruction targets. Although data aggregation offers a potential solution, existing neural topic models often overlook it due to time complexity, poor aggregation quality, and difficulty in inferring topic proportions for individual documents. In this paper, we propose a novel model, GloCOM (Global Clustering COntexts for Topic Models), which addresses these challenges by constructing aggregated global clustering contexts for short documents, leveraging text embeddings from pre-trained language models. GloCOM can infer both global topic distributions for clustering contexts and local distributions for individual short texts. Additionally, the model incorporates these global contexts to augment the reconstruction loss, effectively handling the label sparsity issue. Extensive experiments on short text datasets show that our approach outperforms other state-of-the-art models in both topic quality and document representations.

Tung Nguyen, Aditya Grover

Optimizing black-box functions is a fundamental problem in science and engineering. To solve this problem, many approaches learn a surrogate function that estimates the underlying objective from limited historical evaluations. Large Language Models (LLMs), with their strong pattern-matching capabilities via pretraining on vast amounts of data, stand out as a potential candidate for surrogate modeling. However, directly prompting a pretrained language model to produce predictions is not feasible in many scientific domains due to the scarcity of domain-specific data in the pretraining corpora and the challenges of articulating complex problems in natural language. In this work, we introduce LICO, a general-purpose model that extends arbitrary base LLMs for black-box optimization, with a particular application to the molecular domain. To achieve this, we equip the language model with a separate embedding layer and prediction layer, and train the model to perform in-context predictions on a diverse set of functions defined over the domain. Once trained, LICO can generalize to unseen molecule properties simply via in-context prompting. LICO performs competitively on PMO, a challenging molecular optimization benchmark comprising 23 objective functions, and achieves state-of-the-art performance on its low-budget version PMO-1K.

Siyan Zhao, Tung Nguyen, Aditya Grover

In-context learning is a key paradigm in large language models (LLMs) that enables them to generalize to new tasks and domains by simply prompting these models with a few exemplars without explicit parameter updates. Many attempts have been made to understand in-context learning in LLMs as a function of model scale, pretraining data, and other factors. In this work, we propose a new mechanism to probe and understand in-context learning from the lens of decision boundaries for in-context binary classification. Decision boundaries are straightforward to visualize and provide important information about the qualitative behavior of the inductive biases of standard classifiers. To our surprise, we find that the decision boundaries learned by current LLMs in simple binary classification tasks are often irregular and non-smooth, regardless of linear separability in the underlying task. This paper investigates the factors influencing these decision boundaries and explores methods to enhance their generalizability. We assess various approaches, including training-free and fine-tuning methods for LLMs, the impact of model architecture, and the effectiveness of active prompting techniques for smoothing decision boundaries in a data-efficient manner. Our findings provide a deeper understanding of in-context learning dynamics and offer practical improvements for enhancing robustness and generalizability of in-context learning.

Phong Minh Vu, Tam The Nguyen, Tung Thanh Nguyen

Recent studies showed that the dialogs between app developers and app users on app stores are important to increase user satisfaction and app's overall ratings. However, the large volume of reviews and the limitation of resources discourage app developers from engaging with customers through this channel. One solution to this problem is to develop an Automated Responding System for developers to respond to app reviews in a manner that is most similar to a human response. Toward designing such system, we have conducted an empirical study of the characteristics of mobile apps' reviews and their human-written responses. We found that an app reviews can have multiple fragments at sentence level with different topics and intentions. Similarly, a response also can be divided into multiple fragments with unique intentions to answer certain parts of their review (e.g., complaints, requests, or information seeking). We have also identified several characteristics of review (rating, topics, intentions, quantitative text feature) that can be used to rank review by their priority of need for response. In addition, we identified the degree of re-usability of past responses is based on their context (single app, apps of the same category, and their common features). Last but not least, a responses can be reused in another review if some parts of it can be replaced by a placeholder that is either a named-entity or a hyperlink. Based on those findings, we discuss the implications of developing an Automated Responding System to help mobile apps' developers write the responses for users reviews more effectively.

Tam The Nguyen, Phong Minh Vu, Tung Thanh Nguyen

In modern programming languages, exception handling is an effective mechanism to avoid unexpected runtime errors. Thus, failing to catch and handle exceptions could lead to serious issues like system crashing, resource leaking, or negative end-user experiences. However, writing correct exception handling code is often challenging in mobile app development due to the fast-changing nature of API libraries for mobile apps and the insufficiency of their documentation and source code examples. Our prior study shows that in practice mobile app developers cause many exception-related bugs and still use bad exception handling practices (e.g. catch an exception and do nothing). To address such problems, in this paper, we introduce two novel techniques for recommending correct exception handling code. One technique, XRank, recommends code to catch an exception likely occurring in a code snippet. The other, XHand, recommends correction code for such an occurring exception. We have developed ExAssist, a code recommendation tool for exception handling using XRank and XHand. The empirical evaluation shows that our techniques are highly effective. For example, XRank has top-1 accuracy of 70% and top-3 accuracy of 87%. XHand's results are 89% and 96%, respectively.

Tam The Nguyen, Phong Minh Vu, Tung Thanh Nguyen

Modern software development relies heavily on Application Programming Interface (API) libraries. However, there are often certain constraints on using API elements in such libraries. Failing to follow such constraints (API misuse) could lead to serious programming errors. Many approaches have been proposed to detect API misuses, but they still have low accuracy and cannot repair the detected misuses. In this paper, we propose SAM, a novel approach to detect and repair API misuses automatically. SAM uses statistical models to describe five factors involving in any API method call: related method calls, exceptions, pre-conditions, post-conditions, and values of arguments. These statistical models are trained from a large repository of high-quality production code. Then, given a piece of code, SAM verifies each of its method calls with the trained statistical models. If a factor has a sufficiently low probability, the corresponding call is considered as an API misuse. SAM performs an optimal search for editing operations to apply on the code until it has no API issue.

Hung Pham, Tam Nguyen, Phong Vu et al.

In this paper, we define visual log of a software system as data capturing the interactions between its users and its graphic user interface (GUI), such as screen-shots and screen recordings. We vision that mining such visual log could be useful for bug reproducing and debugging, automated GUI testing, user interface designing, question answering of common usages in software support, etc. Toward that vision, we propose a core framework for mining visual log of software. This framework focuses on detecting GUI elements and changes in visual log, removing users' private data, recognizing user interactions with GUI elements, and learning GUI usage patterns. We also performed a small study on the characteristics of GUI elements in mobile apps. The findings from this study suggested several heuristics to design techniques for recognizing GUI elements and interactions.

Tam The Nguyen, Hung Viet Pham, Phong Minh Vu et al.

When developing mobile apps, programmers rely heavily on standard API frameworks and libraries. However, learning and using those APIs is often challenging due to the fast-changing nature of API frameworks for mobile systems, the complexity of API usages, the insufficiency of documentation, and the unavailability of source code examples. In this paper, we propose a novel approach to learn API usages from bytecode of Android mobile apps. Our core contributions include: i) ARUS, a graph-based representation of API usage scenarios; ii) HAPI, a statistical, generative model of API usages; and iii) three algorithms to extract ARUS from apps' bytecode, to train HAPI based on method call sequences extracted from ARUS, and to recommend method calls in code completion engines using the trained HAPI. Our empirical evaluation suggests that our approach can learn useful API usage models which can provide recommendations with higher levels of accuracy than the baseline n-gram model.

Phong Minh Vu, Tam The Nguyen, Hung Viet Pham et al.

User reviews of mobile apps often contain complaints or suggestions which are valuable for app developers to improve user experience and satisfaction. However, due to the large volume and noisy-nature of those reviews, manually analyzing them for useful opinions is inherently challenging. To address this problem, we propose MARK, a keyword-based framework for semi-automated review analysis. MARK allows an analyst describing his interests in one or some mobile apps by a set of keywords. It then finds and lists the reviews most relevant to those keywords for further analysis. It can also draw the trends over time of those keywords and detect their sudden changes, which might indicate the occurrences of serious issues. To help analysts describe their interests more effectively, MARK can automatically extract keywords from raw reviews and rank them by their associations with negative reviews. In addition, based on a vector-based semantic representation of keywords, MARK can divide a large set of keywords into more cohesive subsets, or suggest keywords similar to the selected ones.