Hoang Nguyen

Duy M. H. Nguyen, Hoang Nguyen, Nghiem T. Diep et al. · eth-zurich

Obtaining large pre-trained models that can be fine-tuned to new tasks with limited annotated samples has remained an open challenge for medical imaging data. While pre-trained deep networks on ImageNet and vision-language foundation models trained on web-scale data are prevailing approaches, their effectiveness on medical tasks is limited due to the significant domain shift between natural and medical images. To bridge this gap, we introduce LVM-Med, the first family of deep networks trained on large-scale medical datasets. We have collected approximately 1.3 million medical images from 55 publicly available datasets, covering a large number of organs and modalities such as CT, MRI, X-ray, and Ultrasound. We benchmark several state-of-the-art self-supervised algorithms on this dataset and propose a novel self-supervised contrastive learning algorithm using a graph-matching formulation. The proposed approach makes three contributions: (i) it integrates prior pair-wise image similarity metrics based on local and global information; (ii) it captures the structural constraints of feature embeddings through a loss function constructed via a combinatorial graph-matching objective; and (iii) it can be trained efficiently end-to-end using modern gradient-estimation techniques for black-box solvers. We thoroughly evaluate the proposed LVM-Med on 15 downstream medical tasks ranging from segmentation and classification to object detection, and both for the in and out-of-distribution settings. LVM-Med empirically outperforms a number of state-of-the-art supervised, self-supervised, and foundation models. For challenging tasks such as Brain Tumor Classification or Diabetic Retinopathy Grading, LVM-Med improves previous vision-language models trained on 1 billion masks by 6-7% while using only a ResNet-50.

Duy M. H. Nguyen, Hoang Nguyen, Mai T. N. Truong et al. · eth-zurich

Collecting large-scale medical datasets with fully annotated samples for training of deep networks is prohibitively expensive, especially for 3D volume data. Recent breakthroughs in self-supervised learning (SSL) offer the ability to overcome the lack of labeled training samples by learning feature representations from unlabeled data. However, most current SSL techniques in the medical field have been designed for either 2D images or 3D volumes. In practice, this restricts the capability to fully leverage unlabeled data from numerous sources, which may include both 2D and 3D data. Additionally, the use of these pre-trained networks is constrained to downstream tasks with compatible data dimensions. In this paper, we propose a novel framework for unsupervised joint learning on 2D and 3D data modalities. Given a set of 2D images or 2D slices extracted from 3D volumes, we construct an SSL task based on a 2D contrastive clustering problem for distinct classes. The 3D volumes are exploited by computing vectored embedding at each slice and then assembling a holistic feature through deformable self-attention mechanisms in Transformer, allowing incorporating long-range dependencies between slices inside 3D volumes. These holistic features are further utilized to define a novel 3D clustering agreement-based SSL task and masking embedding prediction inspired by pre-trained language models. Experiments on downstream tasks, such as 3D brain segmentation, lung nodule detection, 3D heart structures segmentation, and abnormal chest X-ray detection, demonstrate the effectiveness of our joint 2D and 3D SSL approach. We improve plain 2D Deep-ClusterV2 and SwAV by a significant margin and also surpass various modern 2D and 3D SSL approaches.

Matheus Kunzler Maldaner, Raul Valle, Junsung Kim et al.

The growing publication rate of research papers has created an urgent need for better ways to fact-check information, assess writing quality, and identify unverifiable claims. We present Plato's Cave as an open-source, human-centered research verification system that (i) creates a directed acyclic graph (DAG) from a document, (ii) leverages web agents to assign credibility scores to nodes and edges from the DAG, and (iii) gives a final score by interpreting and evaluating the paper's argumentative structure. We report the system implementation and results on a collected dataset of 104 research papers.

Elena Farahbakhsh Touli, Hoang Nguyen, Olha Bodnar

In this paper, we study the connection between the companies in the Swedish capital market. We consider 28 companies included in the determination of the market index OMX30. The network structure of the market is constructed using different methods to determine the distance between the companies. We use hierarchical clustering methods to find the relation among the companies in each window. Next, we obtain one-dimensional time series of the distances between the clustering trees that reflect the changes in the relationship between the companies in the market over time. The method of statistical process control, namely the Shewhart control chart, is applied to those time series to detect abnormal changes in the financial market.

Gemini Team, Petko Georgiev, Ving Ian Lei et al. · deepmind, mila

In this report, we introduce the Gemini 1.5 family of models, representing the next generation of highly compute-efficient multimodal models capable of recalling and reasoning over fine-grained information from millions of tokens of context, including multiple long documents and hours of video and audio. The family includes two new models: (1) an updated Gemini 1.5 Pro, which exceeds the February version on the great majority of capabilities and benchmarks; (2) Gemini 1.5 Flash, a more lightweight variant designed for efficiency with minimal regression in quality. Gemini 1.5 models achieve near-perfect recall on long-context retrieval tasks across modalities, improve the state-of-the-art in long-document QA, long-video QA and long-context ASR, and match or surpass Gemini 1.0 Ultra's state-of-the-art performance across a broad set of benchmarks. Studying the limits of Gemini 1.5's long-context ability, we find continued improvement in next-token prediction and near-perfect retrieval (>99%) up to at least 10M tokens, a generational leap over existing models such as Claude 3.0 (200k) and GPT-4 Turbo (128k). Finally, we highlight real-world use cases, such as Gemini 1.5 collaborating with professionals on completing their tasks achieving 26 to 75% time savings across 10 different job categories, as well as surprising new capabilities of large language models at the frontier; when given a grammar manual for Kalamang, a language with fewer than 200 speakers worldwide, the model learns to translate English to Kalamang at a similar level to a person who learned from the same content.

Thao V. Ha, Hoang Nguyen, Son T. Huynh et al.

In recent years, the occurrence of falls has increased and has had detrimental effects on older adults. Therefore, various machine learning approaches and datasets have been introduced to construct an efficient fall detection algorithm for the social community. This paper studies the fall detection problem based on a large public dataset, namely the UP-Fall Detection Dataset. This dataset was collected from a dozen of volunteers using different sensors and two cameras. We propose several techniques to obtain valuable features from these sensors and cameras and then construct suitable models for the main problem. The experimental results show that our proposed methods can bypass the state-of-the-art methods on this dataset in terms of accuracy, precision, recall, and F1 score.

Gheorghe Comanici, Eric Bieber, Mike Schaekermann et al. · amazon-science, baidu

In this report, we introduce the Gemini 2.X model family: Gemini 2.5 Pro and Gemini 2.5 Flash, as well as our earlier Gemini 2.0 Flash and Flash-Lite models. Gemini 2.5 Pro is our most capable model yet, achieving SoTA performance on frontier coding and reasoning benchmarks. In addition to its incredible coding and reasoning skills, Gemini 2.5 Pro is a thinking model that excels at multimodal understanding and it is now able to process up to 3 hours of video content. Its unique combination of long context, multimodal and reasoning capabilities can be combined to unlock new agentic workflows. Gemini 2.5 Flash provides excellent reasoning abilities at a fraction of the compute and latency requirements and Gemini 2.0 Flash and Flash-Lite provide high performance at low latency and cost. Taken together, the Gemini 2.X model generation spans the full Pareto frontier of model capability vs cost, allowing users to explore the boundaries of what is possible with complex agentic problem solving.

Henry Peng Zou, Wei-Chieh Huang, Yaozu Wu et al.

Recent advances in large language models (LLMs) have sparked growing interest in building fully autonomous agents. However, fully autonomous LLM-based agents still face significant challenges, including limited reliability due to hallucinations, difficulty in handling complex tasks, and substantial safety and ethical risks, all of which limit their feasibility and trustworthiness in real-world applications. To overcome these limitations, LLM-based human-agent systems (LLM-HAS) incorporate human-provided information, feedback, or control into the agent system to enhance system performance, reliability and safety. These human-agent collaboration systems enable humans and LLM-based agents to collaborate effectively by leveraging their complementary strengths. This paper provides the first comprehensive and structured survey of LLM-HAS. It clarifies fundamental concepts, systematically presents core components shaping these systems, including environment & profiling, human feedback, interaction types, orchestration and communication, explores emerging applications, and discusses unique challenges and opportunities arising from human-AI collaboration. By consolidating current knowledge and offering a structured overview, we aim to foster further research and innovation in this rapidly evolving interdisciplinary field. Paper lists and resources are available at https://github.com/HenryPengZou/Awesome-Human-Agent-Collaboration-Interaction-Systems.

Hoang Nguyen, Xiaohao Xu, Xiaonan Huang

Monocular depth foundation models achieve remarkable generalization by learning large-scale semantic priors, but this creates a critical vulnerability: they hallucinate illusory 3D structures from geometrically planar but perceptually ambiguous inputs. We term this failure the 3D Mirage. This paper introduces the first end-to-end framework to probe, quantify, and tame this unquantified safety risk. To probe, we present 3D-Mirage, the first benchmark of real-world illusions (e.g., street art) with precise planar-region annotations and context-restricted crops. To quantify, we propose a Laplacian-based evaluation framework with two metrics: the Deviation Composite Score (DCS) for spurious non-planarity and the Confusion Composite Score (CCS) for contextual instability. To tame this failure, we introduce Grounded Self-Distillation, a parameter-efficient strategy that surgically enforces planarity on illusion ROIs while using a frozen teacher to preserve background knowledge, thus avoiding catastrophic forgetting. Our work provides the essential tools to diagnose and mitigate this phenomenon, urging a necessary shift in MDE evaluation from pixel-wise accuracy to structural and contextual robustness. Our code and benchmark will be publicly available to foster this exciting research direction.

Rishabh Maheshwary, Vikas Yadav, Hoang Nguyen et al.

Instruction finetuning (IFT) is critical for aligning Large Language Models (LLMs) to follow instructions. While many effective IFT datasets have been introduced recently, they predominantly focus on high-resource languages like English. To better align LLMs across a broad spectrum of languages and tasks, we propose a fully synthetic, novel taxonomy (Evol) guided Multilingual, Multi-turn instruction finetuning dataset, called M2Lingual. It is constructed by first selecting a diverse set of seed examples and then utilizing the proposed Evol taxonomy to convert these seeds into complex and challenging multi-turn instructions. We demonstrate the effectiveness of M2Lingual by training LLMs of varying sizes and showcasing the enhanced performance across a diverse set of languages. We contribute the 2 step Evol taxonomy with the guided generation code: https://github.com/ServiceNow/M2Lingual, as well as the first fully synthetic, general and task-oriented, multi-turn, multilingual dataset built with Evol - M2Lingual: https://huggingface.co/datasets/ServiceNow-AI/ M2Lingual - containing 182K total IFT pairs, covering 70 languages and 17+ NLP tasks.

Hoang Nguyen, Xiaohao Xu, Xiaonan Huang

Modern vision-based world models can represent observations as compact yet expressive latent manifolds, but fast goal-oriented planning in these spaces remains challenging. This raises a central question: when does a learned representation simplify control, rather than merely enabling prediction? We study this question in a pretrained LeWorldModel, whose latent geometry is regularized for smoothness and uniformity. Our key insight is that, under such geometry, planning can be amortized into a latent inverse-dynamics mapping instead of requiring online search. We therefore replace iterative planning with a lightweight Goal-Conditioned Inverse Dynamics Model (GC-IDM) that maps the current latent state, goal latent state, and remaining horizon directly to the next action. Empirically, across four benchmark environments spanning navigation, contact-rich manipulation, and continuous control, our controller matches or exceeds CEM in seven of eight environment-protocol settings while reducing per-decision cost by 100-130x. A broader sweep over test-time planners (CEM, MPPI, iCEM, and gradient-based methods) shows that this result is not specific to a particular optimizer. These findings suggest that much of the structure recovered by test-time planning is already locally encoded in the latent representation. More broadly, our results indicate that sufficiently structured latent spaces can shift part of the planning burden from online optimization to learned inference.

Yuri Chervonyi, Trieu H. Trinh, Miroslav Olšák et al.

We present AlphaGeometry2, a significantly improved version of AlphaGeometry introduced in Trinh et al. (2024), which has now surpassed an average gold medalist in solving Olympiad geometry problems. To achieve this, we first extend the original AlphaGeometry language to tackle harder problems involving movements of objects, and problems containing linear equations of angles, ratios, and distances. This, together with support for non-constructive problems, has markedly improved the coverage rate of the AlphaGeometry language on International Math Olympiads (IMO) 2000-2024 geometry problems from 66% to 88%. The search process of AlphaGeometry2 has also been greatly improved through the use of Gemini architecture for better language modeling, and a novel knowledge-sharing mechanism that enables effective communication between search trees. Together with further enhancements to the symbolic engine and synthetic data generation, we have significantly boosted the overall solving rate of AlphaGeometry2 to 84% for $\textit{all}$ geometry problems over the last 25 years, compared to 54% previously. AlphaGeometry2 was also part of the system that achieved silver-medal standard at IMO 2024 https://dpmd.ai/imo-silver. Last but not least, we report progress towards using AlphaGeometry2 as a part of a fully automated system that reliably solves geometry problems directly from natural language input.

An Nguyen, Hoang Nguyen, Phuong Le et al.

We study city-scale control of electric-vehicle (EV) ride-hailing fleets where dispatch, repositioning, and charging decisions must respect charger and feeder limits under uncertain, spatially correlated demand and travel times. We formulate the problem as a hex-grid semi-Markov decision process (semi-MDP) with mixed actions -- discrete actions for serving, repositioning, and charging, together with continuous charging power -- and variable action durations. To guarantee physical feasibility during both training and deployment, the policy learns over high-level intentions produced by a masked, temperature-annealed actor. These intentions are projected at every decision step through a time-limited rolling mixed-integer linear program (MILP) that strictly enforces state-of-charge, port, and feeder constraints. To mitigate distributional shifts, we optimize a Soft Actor--Critic (SAC) agent against a Wasserstein-1 ambiguity set with a graph-aligned Mahalanobis ground metric that captures spatial correlations. The robust backup uses the Kantorovich--Rubinstein dual, a projected subgradient inner loop, and a primal--dual risk-budget update. Our architecture combines a two-layer Graph Convolutional Network (GCN) encoder, twin critics, and a value network that drives the adversary. Experiments on a large-scale EV fleet simulator built from NYC taxi data show that PD--RSAC achieves the highest net profit, reaching \$1.22M, compared with \$0.58M--\$0.70M for strong heuristic, single-agent RL, and multi-agent RL baselines, including Greedy, SAC, MAPPO, and MADDPG, while maintaining zero feeder-limit violations.

Hoang Nguyen, Lu Wang, Marta Gaia Bras

Freight brokerages negotiate thousands of carrier rates daily under dynamic pricing conditions where models frequently revise targets mid-conversation. Classical time-dependent concession frameworks use a fixed shape parameter $β$ that cannot adapt to these updates. Deriving $β$ from the live spread enables adaptation but introduces a new problem: a pricing shift can cause the formula to retract a previous offer, violating monotonicity. LLM-powered brokers offer flexibility but require expensive reasoning models, produce non-deterministic pricing, and remain vulnerable to prompt injection. We propose a two-index anchor-and-resume framework that addresses both limitations. A spread-derived $β$ maps each load's margin structure to the correct concession posture, while the anchor-and-resume mechanism guarantees monotonically non-decreasing offers under arbitrary pricing shifts. All pricing decisions remain in a deterministic formula; the LLM, when used, serves only as a natural-language translation layer. Empirical evaluation across 115,125 negotiations shows that the adaptive $β$ tailors behavior by regime: in narrow spreads, it concedes quickly to prioritize deal closure and load coverage; in medium and wide spreads, it matches or exceeds the best fixed-$β$ baselines in broker savings. Against an unconstrained 20-billion-parameter LLM broker, it achieves similar agreement rates and savings. Against LLM-powered carriers as more realistic stochastic counterparties, it maintains comparable savings and higher agreement rates than against rule-based opponents. By decoupling the LLM from pricing logic, the framework scales horizontally to thousands of concurrent negotiations with negligible inference cost and transparent decision-making.

Hong Nguyen, Hoang Nguyen, Melinda Chang et al.

Understanding the severity of conditions shown in images in medical diagnosis is crucial, serving as a key guide for clinical assessment, treatment, as well as evaluating longitudinal progression. This paper proposes Con- PrO: a novel representation learning method for severity assessment in medical images using Contrastive learningintegrated Preference Optimization. Different from conventional contrastive learning methods that maximize the distance between classes, ConPrO injects into the latent vector the distance preference knowledge between various severity classes and the normal class. We systematically examine the key components of our framework to illuminate how contrastive prediction tasks acquire valuable representations. We show that our representation learning framework offers valuable severity ordering in the feature space while outperforming previous state-of-the-art methods on classification tasks. We achieve a 6% and 20% relative improvement compared to a supervised and a self-supervised baseline, respectively. In addition, we derived discussions on severity indicators and related applications of preference comparison in the medical domain.

Yile Gu, Rohan Kadekodi, Hoang Nguyen et al. · uw

The recent shift in Generative AI (GenAI) applications from cloud-only environments to end-user devices introduces new challenges in resource management, system efficiency, and user experience. This paper presents ConsumerBench, a comprehensive benchmarking framework designed to evaluate the system efficiency and response time of GenAI models running on end-user devices. Unlike existing benchmarks that assume exclusive model access on dedicated GPUs, ConsumerBench simulates realistic multi-application scenarios executing concurrently on constrained hardware. Furthermore, ConsumerBench supports customizable workflows that simulate complex tasks requiring coordination among multiple applications. ConsumerBench captures both application-level metrics, including latency and Service Level Objective (SLO) attainment, and system-level metrics like CPU/GPU utilization and memory bandwidth. Through extensive experiments, ConsumerBench reveals inefficiencies in resource sharing, unfair scheduling under greedy allocation, and performance pitfalls of static model server configurations. The paper also provides practical insights for model developers and system designers, highlighting the benefits of custom kernels tailored to consumer-grade GPU architectures and the value of implementing SLO-aware scheduling strategies.

The Chuong Chu, Vu Tuan Dat Pham, Kien Dao et al.

Intra-sentential code-switching (CS) refers to the alternation between languages that happens within a single utterance and is a significant challenge for Automatic Speech Recognition (ASR) systems. For example, when a Vietnamese speaker uses foreign proper names or specialized terms within their speech. ASR systems often struggle to accurately transcribe intra-sentential CS due to their training on monolingual data and the unpredictable nature of CS. This issue is even more pronounced for low-resource languages, where limited data availability hinders the development of robust models. In this study, we propose AdaCS, a normalization model integrates an adaptive bias attention module (BAM) into encoder-decoder network. This novel approach provides a robust solution to CS ASR in unseen domains, thereby significantly enhancing our contribution to the field. By utilizing BAM to both identify and normalize CS phrases, AdaCS enhances its adaptive capabilities with a biased list of words provided during inference. Our method demonstrates impressive performance and the ability to handle unseen CS phrases across various domains. Experiments show that AdaCS outperforms previous state-of-the-art method on Vietnamese CS ASR normalization by considerable WER reduction of 56.2% and 36.8% on the two proposed test sets.

Sidharth Surapaneni, Hoang Nguyen, Jash Mehta et al.

Large Audio Language Models (LALMs) are rapidly advancing, but evaluating them remains challenging due to inefficient toolkits that limit fair comparison and systematic assessment. Current frameworks suffer from three critical issues: slow processing that bottlenecks large-scale studies, inconsistent prompting that hurts reproducibility, and narrow task coverage that misses important audio reasoning capabilities. We introduce AU-Harness, an efficient and comprehensive evaluation framework for LALMs. Our system achieves a speedup of up to 127% over existing toolkits through optimized batch processing and parallel execution, enabling large-scale evaluations previously impractical. We provide standardized prompting protocols and flexible configurations for fair model comparison across diverse scenarios. Additionally, we introduce two new evaluation categories: LLM-Adaptive Diarization for temporal audio understanding and Spoken Language Reasoning for complex audio-based cognitive tasks. Through evaluation across 380+ tasks, we reveal significant gaps in current LALMs, particularly in temporal understanding and complex spoken language reasoning tasks. Our findings also highlight a lack of standardization in instruction modality existent across audio benchmarks, which can lead up performance differences up to 9.5 absolute points on the challenging complex instruction following downstream tasks. AU-Harness provides both practical evaluation tools and insights into model limitations, advancing systematic LALM development.

Thuy Nguyen, Dang Nguyen, Hoang Nguyen et al.

We present a challenging benchmark for the Open WorLd VISual question answering (OWLViz) task. OWLViz presents concise, unambiguous queries that require integrating multiple capabilities, including visual understanding, web exploration, and specialized tool usage. While humans achieve 69.2% accuracy on these intuitive tasks, even state-of-the-art VLMs struggle, with the best model, Gemini 2.0, achieving only 26.6% accuracy. Current agentic VLMs, which rely on limited vision and vision-language models as tools, perform even worse. This performance gap reveals significant limitations in multimodal systems' ability to select appropriate tools and execute complex reasoning sequences, establishing new directions for advancing practical AI research.

Hoang Nguyen, Chenwei Zhang, Congying Xia et al.

Few-shot Intent Detection is challenging due to the scarcity of available annotated utterances. Although recent works demonstrate that multi-level matching plays an important role in transferring learned knowledge from seen training classes to novel testing classes, they rely on a static similarity measure and overly fine-grained matching components. These limitations inhibit generalizing capability towards Generalized Few-shot Learning settings where both seen and novel classes are co-existent. In this paper, we propose a novel Semantic Matching and Aggregation Network where semantic components are distilled from utterances via multi-head self-attention with additional dynamic regularization constraints. These semantic components capture high-level information, resulting in more effective matching between instances. Our multi-perspective matching method provides a comprehensive matching measure to enhance representations of both labeled and unlabeled instances. We also propose a more challenging evaluation setting that considers classification on the joint all-class label space. Extensive experimental results demonstrate the effectiveness of our method. Our code and data are publicly available.

Congying Xia, Chenwei Zhang, Hoang Nguyen et al.

In this paper, we formulate a more realistic and difficult problem setup for the intent detection task in natural language understanding, namely Generalized Few-Shot Intent Detection (GFSID). GFSID aims to discriminate a joint label space consisting of both existing intents which have enough labeled data and novel intents which only have a few examples for each class. To approach this problem, we propose a novel model, Conditional Text Generation with BERT (CG-BERT). CG-BERT effectively leverages a large pre-trained language model to generate text conditioned on the intent label. By modeling the utterance distribution with variational inference, CG-BERT can generate diverse utterances for the novel intents even with only a few utterances available. Experimental results show that CG-BERT achieves state-of-the-art performance on the GFSID task with 1-shot and 5-shot settings on two real-world datasets.

Girish Kumar, Matthew Henderson, Shannon Chan et al.

Sellers in user to user marketplaces can be inundated with questions from potential buyers. Answers are often already available in the product description. We collected a dataset of around 590K such questions and answers from conversations in an online marketplace. We propose a question answering system that selects a sentence from the product description using a neural-network ranking model. We explore multiple encoding strategies, with recurrent neural networks and feed-forward attention layers yielding good results. This paper presents a demo to interactively pose buyer questions and visualize the ranking scores of product description sentences from live online listings.

Hoang Nguyen, Marlon Dumas, Marcello La Rosa et al.

Business process deviance refers to the phenomenon whereby a subset of the executions of a business process deviate, in a negative or positive way, with respect to its expected or desirable outcomes. Deviant executions of a business process include those that violate compliance rules, or executions that undershoot or exceed performance targets. Deviance mining is concerned with uncovering the reasons for deviant executions by analyzing business process event logs. This article provides a systematic review and comparative evaluation of deviance mining approaches based on a family of data mining techniques known as sequence classification. Using real-life logs from multiple domains, we evaluate a range of feature types and classification methods in terms of their ability to accurately discriminate between normal and deviant executions of a process. We also analyze the interestingness of the rule sets extracted using different methods. We observe that feature sets extracted using pattern mining techniques only slightly outperform simpler feature sets based on counts of individual activity occurrences in a trace.