Guoming Zhang

Aakshita Chandiramani, Aaron Blakeman, Abdullahi Olaoye et al. · amazon-science, cmu

We describe the pre-training, post-training, and quantization of Nemotron 3 Super, a 120 billion (active 12 billion) parameter hybrid Mamba-Attention Mixture-of-Experts model. Nemotron 3 Super is the first model in the Nemotron 3 family to 1) be pre-trained in NVFP4, 2) leverage LatentMoE, a new Mixture-of-Experts architecture that optimizes for both accuracy per FLOP and accuracy per parameter, and 3) include MTP layers for inference acceleration through native speculative decoding. We pre-trained Nemotron 3 Super on 25 trillion tokens followed by post-training using supervised fine tuning (SFT) and reinforcement learning (RL). The final model supports up to 1M context length and achieves comparable accuracy on common benchmarks, while also achieving up to 2.2x and 7.5x higher inference throughput compared to GPT-OSS-120B and Qwen3.5-122B, respectively. Nemotron 3 Super datasets, along with the base, post-trained, and quantized checkpoints, are open-sourced on HuggingFace.

Mingda Han, Huanqi Yang, Zehua Sun et al.

Millimeter-wave (mmWave) radar enables privacy-preserving human activity recognition (HAR), yet real-world deployment remains hindered by costly annotation and poor transferability under domain shift. Although prior efforts partially alleviate these challenges, most still require retraining or adaptation for each new deployment setting. This keeps mmWave HAR in a repeated collect-tune-redeploy cycle, making scalable real-world deployment difficult. In this paper, we present RAGent, a deployment-time training-free framework for mmWave HAR that reformulates recognition as evidence-grounded inference over reusable radar knowledge rather than deployment-specific model optimization. Offline, RAGent constructs a reusable radar knowledge base through constrained cross-modal supervision, where a Vision-Language Model (VLM) transfers activity semantics from synchronized videos to paired radar segments without manual radar annotation. At deployment time, RAGent recognizes activities from radar alone by retrieving physically comparable precedents in an explicit kinematic space and resolving the final label through structured multi-role reasoning. The reasoning protocol is further refined offline through zero-gradient self-evolution. Extensive experiments on a self-collected dataset show that RAGent achieves 93.39% accuracy without per-domain retraining or target-domain adaptation, while generalizing robustly across domains.

Mingda Han, Huanqi Yang, Chaoqun Li et al.

Rapid advances in speech synthesis and audio editing have made realistic forgeries increasingly accessible, yet existing detection methods remain vulnerable to tampering or depend on visual/wearable sensors. In this paper, we present VoxAnchor, a system that physically grounds audio authentication in vocal dynamics by leveraging the inherent coherence between speech acoustics and radar-sensed throat vibrations. VoxAnchor uses contactless millimeter-wave radar to capture fine-grained throat vibrations that are tightly coupled with human speech production, establishing a hard-to-forge anchor rooted in human physiology. The design comprises three main components: (1) a cross-modal frame-work that uses modality-specific encoders and contrastive learning to detect subtle mismatches at word granularity; (2) a phase-aware pipeline that extracts physically consistent, temporally faithful throat vibrations; and (3) a dual-stage strategy that combines signal-level onset detection and semantic-level coherence to align asynchronous radar and audio streams. Unlike liveness detection, which only confirms whether speech occurred, VoxAnchor verifies what was spoken through word-level content consistency, exposing localized edits that preserve identity and global authenticity cues. Extensive evaluations show that VoxAnchor achieves robust, fine-grained detection across diverse forgeries (editing, splicing, replay, deepfake) and conditions, with an overall EER of 0.017, low latency, and modest computational cost.

Jiong Gong, Haihao Shen, Guoming Zhang et al.

High throughput and low latency inference of deep neural networks are critical for the deployment of deep learning applications. This paper presents the efficient inference techniques of IntelCaffe, the first Intel optimized deep learning framework that supports efficient 8-bit low precision inference and model optimization techniques of convolutional neural networks on Intel Xeon Scalable Processors. The 8-bit optimized model is automatically generated with a calibration process from FP32 model without the need of fine-tuning or retraining. We show that the inference throughput and latency with ResNet-50, Inception-v3 and SSD are improved by 1.38X-2.9X and 1.35X-3X respectively with neglectable accuracy loss from IntelCaffe FP32 baseline and by 56X-75X and 26X-37X from BVLC Caffe. All these techniques have been open-sourced on IntelCaffe GitHub1, and the artifact is provided to reproduce the result on Amazon AWS Cloud.

Shuai Zhao, Yulin Zhang, Luwei Xiao et al.

Despite the remarkable progress of large language models (LLMs) across various domains, their capacity to predict retinopathy of prematurity (ROP) risk remains largely unexplored. To address this gap, we introduce a novel Chinese benchmark dataset, termed CROP, comprising 993 admission records annotated with low, medium, and high-risk labels. To systematically examine the predictive capabilities and affective biases of LLMs in ROP risk stratification, we propose Affective-ROPTester, an automated evaluation framework incorporating three prompting strategies: Instruction-based, Chain-of-Thought (CoT), and In-Context Learning (ICL). The Instruction scheme assesses LLMs' intrinsic knowledge and associated biases, whereas the CoT and ICL schemes leverage external medical knowledge to enhance predictive accuracy. Crucially, we integrate emotional elements at the prompt level to investigate how different affective framings influence the model's ability to predict ROP and its bias patterns. Empirical results derived from the CROP dataset yield two principal observations. First, LLMs demonstrate limited efficacy in ROP risk prediction when operating solely on intrinsic knowledge, yet exhibit marked performance gains when augmented with structured external inputs. Second, affective biases are evident in the model outputs, with a consistent inclination toward overestimating medium- and high-risk cases. Third, compared to negative emotions, positive emotional framing contributes to mitigating predictive bias in model outputs. These findings highlight the critical role of affect-sensitive prompt engineering in enhancing diagnostic reliability and emphasize the utility of Affective-ROPTester as a framework for evaluating and mitigating affective bias in clinical language modeling systems.

Jiahao Wang, Sikun Yang, Heinz Koeppl et al.

Probabilistic approaches for handling count-valued time sequences have attracted amounts of research attentions because their ability to infer explainable latent structures and to estimate uncertainties, and thus are especially suitable for dealing with \emph{noisy} and \emph{incomplete} count data. Among these models, Poisson-Gamma Dynamical Systems (PGDSs) are proven to be effective in capturing the evolving dynamics underlying observed count sequences. However, the state-of-the-art PGDS still fails to capture the \emph{time-varying} transition dynamics that are commonly observed in real-world count time sequences. To mitigate this gap, a non-stationary PGDS is proposed to allow the underlying transition matrices to evolve over time, and the evolving transition matrices are modeled by sophisticatedly-designed Dirichlet Markov chains. Leveraging Dirichlet-Multinomial-Beta data augmentation techniques, a fully-conjugate and efficient Gibbs sampler is developed to perform posterior simulation. Experiments show that, in comparison with related models, the proposed non-stationary PGDS achieves improved predictive performance due to its capacity to learn non-stationary dependency structure captured by the time-evolving transition matrices.

Fangtian Zhong, Zekai Chen, Minghui Xu et al.

Malware is a piece of software that was written with the intent of doing harm to data, devices, or people. Since a number of new malware variants can be generated by reusing codes, malware attacks can be easily launched and thus become common in recent years, incurring huge losses in businesses, governments, financial institutes, health providers, etc. To defeat these attacks, malware classification is employed, which plays an essential role in anti-virus products. However, existing works that employ either static analysis or dynamic analysis have major weaknesses in complicated reverse engineering and time-consuming tasks. In this paper, we propose a visualized malware classification framework called VisMal, which provides highly efficient categorization with acceptable accuracy. VisMal converts malware samples into images and then applies a contrast-limited adaptive histogram equalization algorithm to enhance the similarity between malware image regions in the same family. We provided a proof-of-concept implementation and carried out an extensive evaluation to verify the performance of our framework. The evaluation results indicate that VisMal can classify a malware sample within 4.0ms and have an average accuracy of 96.0%. Moreover, VisMal provides security engineers with a simple visualization approach to further validate its performance.

Guoming Zhang, Chen Yan, Xiaoyu Ji et al.

Speech recognition (SR) systems such as Siri or Google Now have become an increasingly popular human-computer interaction method, and have turned various systems into voice controllable systems(VCS). Prior work on attacking VCS shows that the hidden voice commands that are incomprehensible to people can control the systems. Hidden voice commands, though hidden, are nonetheless audible. In this work, we design a completely inaudible attack, DolphinAttack, that modulates voice commands on ultrasonic carriers (e.g., f > 20 kHz) to achieve inaudibility. By leveraging the nonlinearity of the microphone circuits, the modulated low frequency audio commands can be successfully demodulated, recovered, and more importantly interpreted by the speech recognition systems. We validate DolphinAttack on popular speech recognition systems, including Siri, Google Now, Samsung S Voice, Huawei HiVoice, Cortana and Alexa. By injecting a sequence of inaudible voice commands, we show a few proof-of-concept attacks, which include activating Siri to initiate a FaceTime call on iPhone, activating Google Now to switch the phone to the airplane mode, and even manipulating the navigation system in an Audi automobile. We propose hardware and software defense solutions. We validate that it is feasible to detect DolphinAttack by classifying the audios using supported vector machine (SVM), and suggest to re-design voice controllable systems to be resilient to inaudible voice command attacks.