Siyuan He

Siyuan He, Songlin Jia, Yuyan Bao et al.

Static resource management in languages remains challenging due to tensions among control, expressiveness, and flexibility. Region-based systems [Grossman et al . 2002; Tofte et al. 2001] offer bulk deallocation via lexically scoped regions, where all allocations follow a stack discipline. However, both regions and their resources are second-class, and neither can escape its scope nor be freely returned. Ownership and linear type systems, exemplified by Rust [Clarke et al. 2013], offer non-lexical lifetimes and robust static guarantees, but rely on invariants that limit higher-order patterns and expressive sharing. In this work, we propose a new type system that unifies these strengths. Our system treats all heap-allocated resources as first-class values, while allowing programmers to control lifetime and granularity through three allocation modes: (1) fresh allocation for individual, non-lexical references; (2) subsequent coallocation grouping resources collectively within shadow arenas; and (3) scoped allocation with lexically bounded lifetimes following stack discipline. Regardless of mode, all resources share a uniform type and have no distinction for generic abstractions, preserving the higher-order parametric nature of the language. Obtaining static safety in higher-order languages with flexible sharing is nontrivial. We address this by extending reachability types [Wei et al. 2024] to collectively track first-class resources, and by adopting flow-insensitive deallocation reasoning for selective stack discipline. These mechanisms yield Aq<: and {A}q<: atop, both formalized and proven type safe and memory safe in Rocq.

Siyuan He, Songlin Jia, Tiark Rompf

Dynamic languages such as Python and JavaScript widely use function decorators to extend behavior. In TypeScript, a common way to type such patterns uses Parameters<T> and ReturnType<T>. In practice, this idiom relies on a function-type bound for T that is expressed using the unsafe type any, which weakens static guarantees. At the core is a standard typing principle: application is justified only when the callee is exposed as an arrow type. We present F<:DR, a calculus that adds domain and range projection types, Dom(T) and Range(T), for arbitrary types T. These projections permit typing applications through abstract function types: an argument of type Dom(T) witnesses callability, and the result is typed as Range(T). This design complements, rather than replaces, standard arrow-based application, which remains admissible via subtyping in System F<:. We mechanize F<:DR in Rocq and prove semantic type soundness using logical relations with path selection, which delays projection interpretation until function structure is resolved. The same technique extends to additional projection types, illustrated for primitive pairs, i.e., product types.

Ruikai Cui, Siyuan He, Shi Qiu

Foundation models, such as OpenAI's GPT-3 and GPT-4, Meta's LLaMA, and Google's PaLM2, have revolutionized the field of artificial intelligence. A notable paradigm shift has been the advent of the Segment Anything Model (SAM), which has exhibited a remarkable capability to segment real-world objects, trained on 1 billion masks and 11 million images. Although SAM excels in general object segmentation, it lacks the intrinsic ability to detect salient objects, resulting in suboptimal performance in this domain. To address this challenge, we present the Segment Salient Object Model (SSOM), an innovative approach that adaptively fine-tunes SAM for salient object detection by harnessing the low-rank structure inherent in deep learning. Comprehensive qualitative and quantitative evaluations across five challenging RGB benchmark datasets demonstrate the superior performance of our approach, surpassing state-of-the-art methods.

Kai Ma, Siyuan He, Pengcheng Xi et al.

Computer vision and machine learning are playing an increasingly important role in computer-assisted diagnosis; however, the application of deep learning to medical imaging has challenges in data availability and data imbalance, and it is especially important that models for medical imaging are built to be trustworthy. Therefore, we propose TRUDLMIA, a trustworthy deep learning framework for medical image analysis, which adopts a modular design, leverages self-supervised pre-training, and utilizes a novel surrogate loss function. Experimental evaluations indicate that models generated from the framework are both trustworthy and high-performing. It is anticipated that the framework will support researchers and clinicians in advancing the use of deep learning for dealing with public health crises including COVID-19.

Chunyu Wei, Siyuan He, Yu Wang et al.

Graph anomaly detection (GAD) is crucial in applications like fraud detection and cybersecurity. Despite recent advancements using graph neural networks (GNNs), two major challenges persist. At the model level, most methods adopt a transductive learning paradigm, which assumes static graph structures, making them unsuitable for dynamic, evolving networks. At the data level, the extreme class imbalance, where anomalous nodes are rare, leads to biased models that fail to generalize to unseen anomalies. These challenges are interdependent: static transductive frameworks limit effective data augmentation, while imbalance exacerbates model distortion in inductive learning settings. To address these challenges, we propose a novel data-centric framework that integrates dynamic graph modeling with balanced anomaly synthesis. Our framework features: (1) a discrete ego-graph diffusion model, which captures the local topology of anomalies to generate ego-graphs aligned with anomalous structural distribution, and (2) a curriculum anomaly augmentation mechanism, which dynamically adjusts synthetic data generation during training, focusing on underrepresented anomaly patterns to improve detection and generalization. Experiments on five datasets demonstrate that the effectiveness of our framework.

Chunyu Wei, Huaiyu Qin, Siyuan He et al.

Retrieval-Augmented Generation (RAG) has significantly enhanced Large Language Models' ability to access external knowledge, yet current graph-based RAG approaches face two critical limitations in managing hierarchical information: they impose rigid layer-specific compression quotas that damage local graph structures, and they prioritize topological structure while neglecting semantic content. We introduce T-Retriever, a novel framework that reformulates attributed graph retrieval as tree-based retrieval using a semantic and structure-guided encoding tree. Our approach features two key innovations: (1) Adaptive Compression Encoding, which replaces artificial compression quotas with a global optimization strategy that preserves the graph's natural hierarchical organization, and (2) Semantic-Structural Entropy ($S^2$-Entropy), which jointly optimizes for both structural cohesion and semantic consistency when creating hierarchical partitions. Experiments across diverse graph reasoning benchmarks demonstrate that T-Retriever significantly outperforms state-of-the-art RAG methods, providing more coherent and contextually relevant responses to complex queries.

Siyuan He, Pengcheng Xi, Ashkan Ebadi et al.

Effective representation learning is the key in improving model performance for medical image analysis. In training deep learning models, a compromise often must be made between performance and trust, both of which are essential for medical applications. Moreover, models optimized with cross-entropy loss tend to suffer from unwarranted overconfidence in the majority class and over-cautiousness in the minority class. In this work, we integrate a new surrogate loss with self-supervised learning for computer-aided screening of COVID-19 patients using radiography images. In addition, we adopt a new quantification score to measure a model's trustworthiness. Ablation study is conducted for both the performance and the trust on feature learning methods and loss functions. Comparisons show that leveraging the new surrogate loss on self-supervised models can produce label-efficient networks that are both high-performing and trustworthy.

Ping Zhou, Shuo Huang, Qiang Chen et al.

Objective: Skin stretching around the forehead wrinkles is an important method in rhytidectomy. Proper parameters are required to evaluate the surgical effect. In this paper, a simulation method was proposed to obtain the parameters. Methods: Three-dimensional point cloud data with a resolution of 50 μm were employed. First, a smooth supporting contour under the wrinkled forehead was generated via b-spline interpolation and extrapolation to constrain the deformation of the wrinkled zone. Then, based on the vector formed intrinsic finite element (VFIFE) algorithm, the simulation was implemented in Matlab for the deformation of wrinkled forehead skin in the stretching process. Finally, the stress distribution and the residual wrinkles of forehead skin were employed to evaluate the surgical effect. Results: Although the residual wrinkles are similar when forehead wrinkles are finitely stretched, their stress distribution changes greatly. This indicates that the stress distribution in the skin is effective to evaluate the surgical effect, and the forehead wrinkles are easily to be overstretched, which may lead to potential skin injuries. Conclusion: The simulation method can predict stress distribution and residual wrinkles after forehead wrinkle stretching surgery, which can be potentially used to control the surgical process and further reduce risks of skin injury.