arXiv subjects

Haoyuan Liu

Publications and source records attributed to Haoyuan Liu.

10 recordsLinked to original sources

arXiv2026

On-Shell Bootstrap of Loop Inflation Correlators with Spectral Dispersion

We develop a new bootstrap strategy for cosmological correlators at loop level, which we call spectral dispersion. It is based on two conceptual observations that a correlator can be recovered from its on-shell data, also known as nonlocal signals, by analyticity up to local counterterms, and that the on-shell data for a loop process take the form of a discrete sum over quasinormal modes. Technically, our method combines the dS spectral decomposition with dispersion relations. Using this technique, we bootstrap new results in a simple and intuitive form for 3-point and 4-point correlators with 1-loop massive exchanges of scalar and vector bosons, either directly or derivatively coupled. Applications of this bootstrap technique to higher spins and higher-loop banana graphs with dS covariant dispersions but noncovariant couplings are also straightforward.

hep-th↗

arXiv2026

LiveR: Fine-Grained Elasticity via Live Reconfiguration for Model Training

To reduce user costs and maximize cluster utilization, large model training increasingly leverages volatile but inexpensive GPU capacity, such as spot instances and reclaimable resources in shared clusters. Yet, capitalizing on these economic benefits requires jobs to adapt within the short warning windows that many such environments provide. Existing elastic training systems still treat reconfiguration as stop-and-restart: they externalize distributed state through checkpoints, rebuild the distributed runtime on a new topology, and restart training, turning each resize event into a storage-heavy recovery procedure that incurs substantial downtime from checkpoint I/O, process restart, CUDA initialization, and communicator setup. We present LiveR, a live reconfiguration runtime for elastic LLM training that replaces storage-backed restart with a live, bounded-memory handoff between mixed-parallel training worlds. While the current world continues training, LiveR asynchronously prepares the target world, bootstraps newly added workers in isolation to keep heavyweight initialization off the critical path, and streams model state directly over high-bandwidth interconnects while reshaping it online across tensor, pipeline, and data parallel dimensions. Once the target world is ready, LiveR performs a lightweight commit that switches training to the new configuration without stop-and-restart on the live path. We implement LiveR atop Megatron-LM and PyTorch and evaluate it end-to-end on a multi-node GPU cluster. Across diverse reconfiguration scenarios, LiveR reduces downtime from minutes to seconds, accelerates reconfiguration by 14$\times$-23$\times$ over checkpoint/restart baselines, incurs minimal steady-state overhead, and sustains up to 99% training goodput under volatile-resource conditions, making volatile low-cost GPU capacity far more practical for LLM training.

cs.DC↗

arXiv2026

GVCC: Zero-Shot Video Compression via Codebook-Driven Stochastic Rectified Flow

At ultra-low bitrates, high-fidelity reconstruction requires sampling plausible videos from the posterior rather than regressing to oversmoothed conditional means. We propose Generative Video Codebook Codec (GVCC), a zero-shot framework in which a pretrained video generative model serves directly as the decoder, and the transmitted bitstream specifies its generation trajectory. Modern rectified-flow video models are typically sampled with deterministic ODE solvers, which leave no per-step stochastic channel for transmitting compressed information. GVCC addresses this by converting the deterministic flow sampler into an equivalent marginal-preserving stochastic process, so that information can be transmitted by encoding the per-step stochastic innovations. Unlike images, videos introduce longer temporal dependencies and more diverse conditioning modes. We instantiate GVCC in three practical modes: Text-to-Video (T2V) without a reference frame, autoregressive Image-to-Video (I2V) with tail latent correction, and First-Last-Frame-to-Video (FLF2V) with boundary-sharing Group of Pictures (GOP) chaining. On UVG, GVCC achieves the lowest LPIPS among evaluated baselines across three representative bitrate regimes (down to ${\sim}$0.003\,bpp), with 65\% LPIPS reduction over DCVC-RT at matched bitrate.

cs.CV↗

arXiv2026

FEPLB: Exploiting Copy Engines for Nearly Free MoE Load Balancing in Distributed Training

Fine-grained, per-micro-batch load balancing is essential for efficient Mixture-of-Experts (MoE) training, yet every prior dynamic scheduling scheme pays for it with extra communication that is hard to hide. Especially on modern bulk-transfer backends such as DeepEP. We make a simple but consequential observation: on the NVIDIA Hopper architecture the NVLink Copy Engine can move data between intra-node GPUs without consuming any SM cycles, effectively providing a nearly free communication channel that runs in parallel with compute kernels. FEPLB turns this idle hardware into a new parallel dimension for MoE load rebalancing. Its Two-Phase Dispatch first routes tokens across nodes via the standard EP backend, then redistributes dynamic-expert tokens and weights within the NVLink domain through the Copy Engine at nearly zero cost, while a lightweight CPU scheduler runs concurrently with static expert computation. Because FEPLB uses only Copy Engine and CPU that are orthogonal to those consumed by EP and PP, it coexists with existing parallel strategies without reconfiguration. On GLM-5's MoE layers (128 experts, no auxiliary loss, up to 16 H100 GPUs), FEPLB reduces the token straggler by 51-70% and the GEMM straggler by 50-68% with no measurable EP communication overhead. Its advantage grows with the EP degree: at EP=8, it achieves 2x lower token straggler than FasterMoE.

cs.DC↗

arXiv2025

InterpIoU: Rethinking Bounding Box Regression with Interpolation-Based IoU Optimization

Bounding box regression (BBR) is fundamental to object detection, where the regression loss is crucial for accurate localization. Existing IoU-based losses often incorporate handcrafted geometric penalties to address IoU's non-differentiability in non-overlapping cases and enhance BBR performance. However, these penalties are sensitive to box shape, size, and distribution, often leading to suboptimal optimization for small objects and undesired behaviors such as bounding box enlargement due to misalignment with the IoU objective. To address these limitations, we propose InterpIoU, a novel loss function that replaces handcrafted geometric penalties with a term based on the IoU between interpolated boxes and the target. By using interpolated boxes to bridge the gap between predictions and ground truth, InterpIoU provides meaningful gradients in non-overlapping cases and inherently avoids the box enlargement issue caused by misaligned penalties. Simulation results further show that IoU itself serves as an ideal regression target, while existing geometric penalties are both unnecessary and suboptimal. Building on InterpIoU, we introduce Dynamic InterpIoU, which dynamically adjusts interpolation coefficients based on IoU values, enhancing adaptability to scenarios with diverse object distributions. Experiments on COCO, VisDrone, and PASCAL VOC show that our methods consistently outperform state-of-the-art IoU-based losses across various detection frameworks, with particularly notable improvements in small object detection, confirming their effectiveness.

cs.CV↗

arXiv2025

Enabling Efficient Attack Investigation via Human-in-the-Loop Security Analysis

System auditing is a vital technique for collecting system call events as system provenance and investigating complex multi-step attacks such as Advanced Persistent Threats. However, existing attack investigation methods struggle to uncover long attack sequences due to the massive volume of system provenance data and their inability to focus on attack-relevant parts. In this paper, we present Provexa, a defense system that enables human analysts to effectively analyze large-scale system provenance to reveal multi-step attack sequences. Provexa introduces an expressive domain-specific language, ProvQL, that offers essential primitives for various types of attack analyses (e.g., attack pattern search, attack dependency tracking) with user-defined constraints, enabling analysts to focus on attack-relevant parts and iteratively sift through the large provenance data. Moreover, Provexa provides an optimized execution engine for efficient language execution. Our extensive evaluations on a wide range of attack scenarios demonstrate the practical effectiveness of Provexa in facilitating timely attack investigation.

cs.CR↗

arXiv2024

Massive Inflationary Amplitudes: Differential Equations and Complete Solutions for General Trees

We construct and solve a complete system of differential equations for general tree-level inflation correlators with an arbitrary number of massive scalar exchanges and time-dependent couplings. Any massive tree correlators can be uniquely fixed by solving this system of equations with appropriate boundary conditions. We take a hybrid approach to solve this system, using the differential equation to get the inhomogeneous solution and the bulk time integrals to determine the homogeneous solution. Altogether, we obtain analytical results for all tree-level massive inflation correlators with generic kinematics, expressed as multivariate hypergeometric series of energy ratios. The result can be neatly organized as a sum of the completely inhomogeneous solution, which we call the massive family tree, and all of its cuts. As simple applications, we provide full analytical expressions for tree correlators with one, two, and three massive exchanges.

hep-th↗

arXiv2024

Time Step Generating: A Universal Synthesized Deepfake Image Detector

Currently, high-fidelity text-to-image models are developed in an accelerating pace. Among them, Diffusion Models have led to a remarkable improvement in the quality of image generation, making it vary challenging to distinguish between real and synthesized images. It simultaneously raises serious concerns regarding privacy and security. Some methods are proposed to distinguish the diffusion model generated images through reconstructing. However, the inversion and denoising processes are time-consuming and heavily reliant on the pre-trained generative model. Consequently, if the pre-trained generative model meet the problem of out-of-domain, the detection performance declines. To address this issue, we propose a universal synthetic image detector Time Step Generating (TSG), which does not rely on pre-trained models' reconstructing ability, specific datasets, or sampling algorithms. Our method utilizes a pre-trained diffusion model's network as a feature extractor to capture fine-grained details, focusing on the subtle differences between real and synthetic images. By controlling the time step t of the network input, we can effectively extract these distinguishing detail features. Then, those features can be passed through a classifier (i.e. Resnet), which efficiently detects whether an image is synthetic or real. We test the proposed TSG on the large-scale GenImage benchmark and it achieves significant improvements in both accuracy and generalizability.

cs.CV↗

arXiv2024

Dispersive Bootstrap of Massive Inflation Correlators

Inflation correlators with massive exchanges are central observables of cosmological collider physics, and are also important theoretical data for us to better understand quantum field theories in dS. However, they are difficult to compute directly due to many technical complications of the Schwinger-Keldysh integral. In this work, we initiate a new bootstrap program for massive inflation correlators with dispersion relations on complex momentum planes. We classify kinematic variables of a correlator into vertex energies and line energies, and develop two distinct types of dispersion relations for both of them, respectively called vertex dispersion and line dispersion relations. These dispersion methods allow us to obtain full analytical results of massive correlators from a knowledge of their oscillatory signals alone, while the oscillatory signal at the tree level can be related to simpler subgraphs via the cutting rule. We further apply this method to massive loop correlators, and obtain new analytical expressions for loop diagrams much simpler than existing results from spectral decomposition. In particular, we show that the analyticity demands the existence of an "irreducible background" in the loop correlator, which is unambiguously defined, free of UV divergence, and independent of renormalization schemes.

hep-th↗

arXiv2021

A System for Efficiently Hunting for Cyber Threats in Computer Systems Using Threat Intelligence

Log-based cyber threat hunting has emerged as an important solution to counter sophisticated cyber attacks. However, existing approaches require non-trivial efforts of manual query construction and have overlooked the rich external knowledge about threat behaviors provided by open-source Cyber Threat Intelligence (OSCTI). To bridge the gap, we build ThreatRaptor, a system that facilitates cyber threat hunting in computer systems using OSCTI. Built upon mature system auditing frameworks, ThreatRaptor provides (1) an unsupervised, light-weight, and accurate NLP pipeline that extracts structured threat behaviors from unstructured OSCTI text, (2) a concise and expressive domain-specific query language, TBQL, to hunt for malicious system activities, (3) a query synthesis mechanism that automatically synthesizes a TBQL query from the extracted threat behaviors, and (4) an efficient query execution engine to search the big system audit logging data.

cs.CR↗