Chen Xie

Chen Xie, Xiaodong Gu, Yuling Shi et al.

Code complexity metrics such as cyclomatic complexity have long been used to assess software quality and maintainability. With the rapid advancement of large language models (LLMs) on coding tasks, an important yet underexplored question arises: do traditional complexity metrics meaningfully characterize the coding difficulty that LLMs perceive? In this work, we empirically demonstrate that classical complexity metrics exhibit no consistent correlation with LLM performance, revealing a fundamental mismatch with model-perceived difficulty. To address this gap, we propose LM-CC, a novel code complexity metric tailored for LLMs, grounded in the hypothesis that model-perceived code difficulty is fundamentally driven by semantic nonlinearity. LM-CC quantifies complexity through an entropy-guided semantic compositional hierarchy, capturing the cumulative uncertainty encountered by LLMs during code understanding. Our experimental results demonstrate that LM-CC exhibits strong and consistent partial correlations with LLM performance, while semantics-preserving reductions in LM-CC consistently lead to improved downstream task performance. The source code is available at: https://github.com/xchen121/lm-cc.

Francesco Daghero, Alessio Burrello, Chen Xie et al.

Random Forests (RFs) are widely used Machine Learning models in low-power embedded devices, due to their hardware friendly operation and high accuracy on practically relevant tasks. The accuracy of a RF often increases with the number of internal weak learners (decision trees), but at the cost of a proportional increase in inference latency and energy consumption. Such costs can be mitigated considering that, in most applications, inputs are not all equally difficult to classify. Therefore, a large RF is often necessary only for (few) hard inputs, and wasteful for easier ones. In this work, we propose an early-stopping mechanism for RFs, which terminates the inference as soon as a high-enough classification confidence is reached, reducing the number of weak learners executed for easy inputs. The early-stopping confidence threshold can be controlled at runtime, in order to favor either energy saving or accuracy. We apply our method to three different embedded classification tasks, on a single-core RISC-V microcontroller, achieving an energy reduction from 38% to more than 90% with a drop of less than 0.5% in accuracy. We also show that our approach outperforms previous adaptive ML methods for RFs.

Francesco Daghero, Alessio Burrello, Chen Xie et al.

Human Activity Recognition (HAR) based on inertial data is an increasingly diffused task on embedded devices, from smartphones to ultra low-power sensors. Due to the high computational complexity of deep learning models, most embedded HAR systems are based on simple and not-so-accurate classic machine learning algorithms. This work bridges the gap between on-device HAR and deep learning, proposing a set of efficient one-dimensional Convolutional Neural Networks (CNNs) deployable on general purpose microcontrollers (MCUs). Our CNNs are obtained combining hyper-parameters optimization with sub-byte and mixed-precision quantization, to find good trade-offs between classification results and memory occupation. Moreover, we also leverage adaptive inference as an orthogonal optimization to tune the inference complexity at runtime based on the processed input, hence producing a more flexible HAR system. With experiments on four datasets, and targeting an ultra-low-power RISC-V MCU, we show that (i) We are able to obtain a rich set of Pareto-optimal CNNs for HAR, spanning more than 1 order of magnitude in terms of memory, latency and energy consumption; (ii) Thanks to adaptive inference, we can derive >20 runtime operating modes starting from a single CNN, differing by up to 10% in classification scores and by more than 3x in inference complexity, with a limited memory overhead; (iii) on three of the four benchmarks, we outperform all previous deep learning methods, reducing the memory occupation by more than 100x. The few methods that obtain better performance (both shallow and deep) are not compatible with MCU deployment. (iv) All our CNNs are compatible with real-time on-device HAR with an inference latency <16ms. Their memory occupation varies in 0.05-23.17 kB, and their energy consumption in 0.005 and 61.59 uJ, allowing years of continuous operation on a small battery supply.

Francesco Daghero, Chen Xie, Daniele Jahier Pagliari et al.

Human Activity Recognition (HAR) is a relevant inference task in many mobile applications. State-of-the-art HAR at the edge is typically achieved with lightweight machine learning models such as decision trees and Random Forests (RFs), whereas deep learning is less common due to its high computational complexity. In this work, we propose a novel implementation of HAR based on deep neural networks, and precisely on Binary Neural Networks (BNNs), targeting low-power general purpose processors with a RISC-V instruction set. BNNs yield very small memory footprints and low inference complexity, thanks to the replacement of arithmetic operations with bit-wise ones. However, existing BNN implementations on general purpose processors impose constraints tailored to complex computer vision tasks, which result in over-parametrized models for simpler problems like HAR. Therefore, we also introduce a new BNN inference library, which targets ultra-compact models explicitly. With experiments on a single-core RISC-V processor, we show that BNNs trained on two HAR datasets obtain higher classification accuracy compared to a state-of-the-art baseline based on RFs. Furthermore, our BNN reaches the same accuracy of a RF with either less memory (up to 91%) or more energy-efficiency (up to 70%), depending on the complexity of the features extracted by the RF.

Chen Xie, Francesco Daghero, Yukai Chen et al.

Ultra-low-resolution Infrared (IR) array sensors offer a low-cost, energy-efficient, and privacy-preserving solution for people counting, with applications such as occupancy monitoring. Previous work has shown that Deep Learning (DL) can yield superior performance on this task. However, the literature was missing an extensive comparative analysis of various efficient DL architectures for IR array-based people counting, that considers not only their accuracy, but also the cost of deploying them on memory- and energy-constrained Internet of Things (IoT) edge nodes. In this work, we address this need by comparing 6 different DL architectures on a novel dataset composed of IR images collected from a commercial 8x8 array, which we made openly available. With a wide architectural exploration of each model type, we obtain a rich set of Pareto-optimal solutions, spanning cross-validated balanced accuracy scores in the 55.70-82.70% range. When deployed on a commercial Microcontroller (MCU) by STMicroelectronics, the STM32L4A6ZG, these models occupy 0.41-9.28kB of memory, and require 1.10-7.74ms per inference, while consuming 17.18-120.43 $μ$J of energy. Our models are significantly more accurate than a previous deterministic method (up to +39.9%), while being up to 3.53x faster and more energy efficient. Further, our models' accuracy is comparable to state-of-the-art DL solutions on similar resolution sensors, despite a much lower complexity. All our models enable continuous, real-time inference on a MCU-based IoT node, with years of autonomous operation without battery recharging.

Chen Xie, Francesco Daghero, Yukai Chen et al.

Low-resolution infrared (IR) array sensors offer a low-cost, low-power, and privacy-preserving alternative to optical cameras and smartphones/wearables for social distance monitoring in indoor spaces, permitting the recognition of basic shapes, without revealing the personal details of individuals. In this work, we demonstrate that an accurate detection of social distance violations can be achieved processing the raw output of a 8x8 IR array sensor with a small-sized Convolutional Neural Network (CNN). Furthermore, the CNN can be executed directly on a Microcontroller (MCU)-based sensor node. With results on a newly collected open dataset, we show that our best CNN achieves 86.3% balanced accuracy, significantly outperforming the 61% achieved by a state-of-the-art deterministic algorithm. Changing the architectural parameters of the CNN, we obtain a rich Pareto set of models, spanning 70.5-86.3% accuracy and 0.18-75k parameters. Deployed on a STM32L476RG MCU, these models have a latency of 0.73-5.33ms, with an energy consumption per inference of 9.38-68.57μJ.

Chen Xie, Daniele Jahier Pagliari, Andrea Calimera

Low-resolution infrared (IR) Sensors combined with machine learning (ML) can be leveraged to implement privacy-preserving social distance monitoring solutions in indoor spaces. However, the need of executing these applications on Internet of Things (IoT) edge nodes makes energy consumption critical. In this work, we propose an energy-efficient adaptive inference solution consisting of the cascade of a simple wake-up trigger and a 8-bit quantized Convolutional Neural Network (CNN), which is only invoked for difficult-to-classify frames. Deploying such adaptive system on a IoT Microcontroller, we show that, when processing the output of a 8x8 low-resolution IR sensor, we are able to reduce the energy consumption by 37-57% with respect to a static CNN-based approach, with an accuracy drop of less than 2% (83% balanced accuracy).

Chen Xie, Ciyun Lin, Xiaoyu Zheng et al.

Understanding and predicting pedestrian crossing behavioral intention is crucial for the driving safety of autonomous vehicles. Nonetheless, challenges emerge when using promising images or environmental context masks to extract various factors for time-series network modeling, causing pre-processing errors or a loss of efficiency. Typically, pedestrian positions captured by onboard cameras are often distorted and do not accurately reflect their actual movements. To address these issues, GTransPDM -- a Graph-embedded Transformer with a Position Decoupling Module -- was developed for pedestrian crossing intention prediction by leveraging multi-modal features. First, a positional decoupling module was proposed to decompose pedestrian lateral motion and encode depth cues in the image view. Then, a graph-embedded Transformer was designed to capture the spatio-temporal dynamics of human pose skeletons, integrating essential factors such as position, skeleton, and ego-vehicle motion. Experimental results indicate that the proposed method achieves 92% accuracy on the PIE dataset and 87% accuracy on the JAAD dataset, with a processing speed of 0.05ms. It outperforms the state-of-the-art in comparison.

Junkai Xia, Chenxin Xu, Qingyao Xu et al.

Realistic trajectory generation with natural language control is pivotal for advancing autonomous vehicle technology. However, previous methods focus on individual traffic participant trajectory generation, thus failing to account for the complexity of interactive traffic dynamics. In this work, we propose InteractTraj, the first language-driven traffic trajectory generator that can generate interactive traffic trajectories. InteractTraj interprets abstract trajectory descriptions into concrete formatted interaction-aware numerical codes and learns a mapping between these formatted codes and the final interactive trajectories. To interpret language descriptions, we propose a language-to-code encoder with a novel interaction-aware encoding strategy. To produce interactive traffic trajectories, we propose a code-to-trajectory decoder with interaction-aware feature aggregation that synergizes vehicle interactions with the environmental map and the vehicle moves. Extensive experiments show our method demonstrates superior performance over previous SoTA methods, offering a more realistic generation of interactive traffic trajectories with high controllability via diverse natural language commands. Our code is available at https://github.com/X1a-jk/InteractTraj.git

Jasper Lu, Zhenhao Shen, Yuanfei Wang et al.

Learning robust robot policies in real-world environments requires diverse data augmentation, yet scaling real-world data collection is costly due to the need for acquiring physical assets and reconfiguring environments. Therefore, augmenting real-world scenes into simulation has become a practical augmentation for efficient learning and evaluation. We present a generative framework that establishes a generative real-to-sim mapping from real-world panoramas to high-fidelity simulation scenes, and further synthesize diverse cousin scenes via semantic and geometric editing. Combined with high-quality physics engines and realistic assets, the generated scenes support interactive manipulation tasks. Additionally, we incorporate multi-room stitching to construct consistent large-scale environments for long-horizon navigation across complex layouts. Experiments demonstrate a strong sim-to-real correlation validating our platform's fidelity, and show that extensively scaling up data generation leads to significantly better generalization to unseen scene and object variations, demonstrating the effectiveness of Digital Cousins for generalizable robot learning and evaluation.

Feng Jiang, Yang Chen, Kyle Xu et al.

Recent advances in large-scale video world models have enabled increasingly realistic future prediction, raising the prospect of leveraging imagined videos for robot learning. However, visual realism does not imply physical plausibility, and behaviors inferred from generated videos may violate dynamics and fail when executed by embodied agents. Existing benchmarks begin to incorporate notions of physical plausibility, but they largely remain perception- or diagnostic-oriented and do not systematically evaluate whether predicted behaviors can be translated into executable actions that complete the intended task. To address this gap, we introduce RoboWM-Bench, a manipulation-centric benchmark for embodiment-grounded evaluation of video world models. RoboWM-Bench converts generated behaviors from both human-hand and robotic manipulation videos into embodied action sequences and validates them through robotic execution. The benchmark spans diverse manipulation scenarios and establishes a unified protocol for consistent and reproducible evaluation. Using RoboWM-Bench, we evaluate state-of-the-art video world models and find that reliably generating physically executable behaviors remains an open challenge. Common failure modes include errors in spatial reasoning, unstable contact prediction, and non-physical deformations. While finetuning on manipulation data yields improvements, physical inconsistencies still persist, suggesting opportunities for more physically grounded video generation for robots.

Haodi Yao, Fenghua He, Ning Hao et al.

The field of keypoint extraction, which is essential for vision applications like Structure from Motion (SfM) and Simultaneous Localization and Mapping (SLAM), has evolved from relying on handcrafted methods to leveraging deep learning techniques. While deep learning approaches have significantly improved performance, they often incur substantial computational costs, limiting their deployment in real-time edge applications. Efforts to create lightweight neural networks have seen some success, yet they often result in trade-offs between efficiency and accuracy. Additionally, the high-dimensional descriptors generated by these networks poses challenges for distributed applications requiring efficient communication and coordination, highlighting the need for compact yet competitively accurate descriptors. In this paper, we present EdgePoint2, a series of lightweight keypoint detection and description neural networks specifically tailored for edge computing applications on embedded system. The network architecture is optimized for efficiency without sacrificing accuracy. To train compact descriptors, we introduce a combination of Orthogonal Procrustes loss and similarity loss, which can serve as a general approach for hypersphere embedding distillation tasks. Additionally, we offer 14 sub-models to satisfy diverse application requirements. Our experiments demonstrate that EdgePoint2 consistently achieves state-of-the-art (SOTA) accuracy and efficiency across various challenging scenarios while employing lower-dimensional descriptors (32/48/64). Beyond its accuracy, EdgePoint2 offers significant advantages in flexibility, robustness, and versatility. Consequently, EdgePoint2 emerges as a highly competitive option for visual tasks, especially in contexts demanding adaptability to diverse computational and communication constraints.

Chen Xie, Mingsheng Jiao, Xiaodong Gu et al.

While large language models (LLMs) have been widely applied to code generation, they struggle with generating entire deep learning projects, which are characterized by complex structures, longer functions, and stronger reliance on domain knowledge than general-purpose code. An open-domain LLM often lacks coherent contextual guidance and domain expertise for specific projects, making it challenging to produce complete code that fully meets user requirements. In this paper, we propose a novel planning-guided code generation method, DLCodeGen, tailored for generating deep learning projects. DLCodeGen predicts a structured solution plan, offering global guidance for LLMs to generate the project. The generated plan is then leveraged to retrieve semantically analogous code samples and subsequently abstract a code template. To effectively integrate these multiple retrieval-augmented techniques, a comparative learning mechanism is designed to generate the final code. We validate the effectiveness of our approach on a dataset we build for deep learning code generation. Experimental results demonstrate that DLCodeGen outperforms other baselines, achieving improvements of 9.7% in CodeBLEU and 3.6% in human evaluation metrics.

Matteo Risso, Chen Xie, Francesco Daghero et al.

Low-resolution infrared (IR) array sensors enable people counting applications such as monitoring the occupancy of spaces and people flows while preserving privacy and minimizing energy consumption. Deep Neural Networks (DNNs) have been shown to be well-suited to process these sensor data in an accurate and efficient manner. Nevertheless, the space of DNNs' architectures is huge and its manual exploration is burdensome and often leads to sub-optimal solutions. To overcome this problem, in this work, we propose a highly automated full-stack optimization flow for DNNs that goes from neural architecture search, mixed-precision quantization, and post-processing, down to the realization of a new smart sensor prototype, including a Microcontroller with a customized instruction set. Integrating these cross-layer optimizations, we obtain a large set of Pareto-optimal solutions in the 3D-space of energy, memory, and accuracy. Deploying such solutions on our hardware platform, we improve the state-of-the-art achieving up to 4.2x model size reduction, 23.8x code size reduction, and 15.38x energy reduction at iso-accuracy.

Xingzu Zhan, Chen Xie, Honghang Chen et al.

Text-to-motion generation, which converts motion language descriptions into coherent 3D human motion sequences, has attracted increasing attention in fields, such as avatar animation and humanoid robotic interaction. Though existing models have achieved significant fidelity, they still suffer from two core limitations: (i) They treat motion periodicity and keyframe saliency as independent factors, overlooking their coupling and causing generation drift in long sequences. (ii) They are fragile to semantically equivalent paraphrases, where minor synonym substitutions distort textual embeddings, propagating through the decoder and producing unstable or erroneous motions. In this work, we propose T2M Mamba to address these limitations by (i) proposing Periodicity-Saliency Aware Mamba, which utilizes novel algorithms for keyframe weight estimation via enhanced Density Peaks Clustering and motion periodicity estimation via FFT-accelerated autocorrelation to capture coupled dynamics with minimal computational overhead, and (ii) constructing a Periodic Differential Cross-modal Alignment Module (PDCAM) to enhance robust alignment of textual and motion embeddings. Extensive experiments on HumanML3D and KIT-ML datasets have been conducted, confirming the effectiveness of our approach, achieving an FID of 0.068 and consistent gains on all other metrics.

Di Zhu, Chen Xie, Ziwei Wang et al.

New York City reports over one hundred thousand motor vehicle collisions each year, creating substantial injury and public health burden. We present RaX-Crash, a resource efficient and explainable small model pipeline for structured injury severity prediction on the official NYC Motor Vehicle Collisions dataset. RaX-Crash integrates three linked tables with tens of millions of records, builds a unified feature schema in partitioned storage, and trains compact tree based ensembles (Random Forest and XGBoost) on engineered tabular features, which are compared against locally deployed small language models (SLMs) prompted with textual summaries. On a temporally held out test set, XGBoost and Random Forest achieve accuracies of 0.7828 and 0.7794, clearly outperforming SLMs (0.594 and 0.496); class imbalance analysis shows that simple class weighting improves fatal recall with modest accuracy trade offs, and SHAP attribution highlights human vulnerability factors, timing, and location as dominant drivers of predicted severity. Overall, RaX-Crash indicates that interpretable small model ensembles remain strong baselines for city scale injury analytics, while hybrid pipelines that pair tabular predictors with SLM generated narratives improve communication without sacrificing scalability.

Chen Xie, Gaofeng Wu, Kaidong Wang et al.

Empathy in young children is crucial for their social and emotional development, yet predicting it remains challenging. Traditional methods often only rely on self-reports or observer-based labeling, which are susceptible to bias and fail to objectively capture the process of empathy formation. EEG offers an objective alternative; however, current approaches primarily extract static patterns, neglecting temporal dynamics. To overcome these limitations, we propose a novel deep learning framework, the Brainwave Empathy Assessment Model (BEAM), to predict empathy levels in children aged 4-6 years. BEAM leverages multi-view EEG signals to capture both cognitive and emotional dimensions of empathy. The framework comprises three key components: 1) a LaBraM-based encoder for effective spatio-temporal feature extraction, 2) a feature fusion module to integrate complementary information from multi-view signals, and 3) a contrastive learning module to enhance class separation. Validated on the CBCP dataset, BEAM outperforms state-of-the-art methods across multiple metrics, demonstrating its potential for objective empathy assessment and providing a preliminary insight into early interventions in children's prosocial development.

Xingzu Zhan, Chen Xie, Honghang Chen et al.

Text-to-motion generation sits at the intersection of multimodal learning and computer graphics and is gaining momentum because it can simplify content creation for games, animation, robotics and virtual reality. Most current methods stack spatial and temporal features in a straightforward way, which adds redundancy and still misses subtle joint-level cues. We introduce HiSTF Mamba, a framework with three parts: Dual-Spatial Mamba, Bi-Temporal Mamba and a Dynamic Spatiotemporal Fusion Module (DSFM). The Dual-Spatial module runs part-based and whole-body models in parallel, capturing both overall coordination and fine-grained joint motion. The Bi-Temporal module scans sequences forward and backward to encode short-term details and long-term dependencies. DSFM removes redundant temporal information, extracts complementary cues and fuses them with spatial features to build a richer spatiotemporal representation. Experiments on the HumanML3D benchmark show that HiSTF Mamba performs well across several metrics, achieving high fidelity and tight semantic alignment between text and motion.