arXiv subjects

Yupeng He

Publications and source records attributed to Yupeng He.

7 recordsLinked to original sources

arXiv2026

Navigating the Open-Source Model Ecosystem: An Empirical Study of Creator Practices in Artistic Image Generation

The open-sourcing of powerful image generation models has created a vibrant ecosystem where creators curate and combine a vast array of community-contributed models. This practice stands in sharp contrast to using closed-source tools like Midjourney. Yet, little is known about these emerging creative workflows. To bridge this gap, this paper presents the first large-scale empirical study of creator model usage behavior within this open-source image generation ecosystem. We construct a novel dataset of 6 million images with their embedded generation metadata -- a detailed recipe of the creation process, including the models used and the prompts. By linking the usage of 22.4K base models and 154K LoRA models to the images, our findings underscore the ecosystem's unique strengths and its inherent obstacles. This provides valuable insights for making this ecosystem more sustainable and innovative. Moreover, we make our dataset publicly available, providing creators with practical references for producing better artworks and researchers to facilitate further studies.

cs.HC↗

arXiv2026

Behind EvoMap: Characterizing a Self-Evolving Agent-to-Agent Collaboration Network

Agent-to-Agent (A2A) networks enable autonomous AI agents to collaborate by sharing reusable problem-solving instructions. However, how these decentralized ecosystems operate in practice remains largely unexplored. We present the first large-scale empirical study of EvoMap, a prominent A2A collaboration network. By analyzing over 1.5M assets and 128K agents, we show how design choices that prioritize scalable growth introduce trade-offs in reusability, evolution, and auditability. First, EvoMap's credit economy rewards agents for publishing valuable assets. Although this design encourages participation at scale, rewards are tied primarily to publication rather than adoption. This leads agents to mass-produce assets to accumulate credits. As a result, 98% of assets are never reused, while rewards become highly concentrated among a small fraction of agents. Second, EvoMap employs an algorithm (referred to as GDI) to score and rank the quality of these shared assets. We demonstrate that this scoring system is flawed: rather than measuring objective performance, an asset's rank is heavily dictated by unverified, self-reported metadata (e.g., claimed lines of code modified). This allows agents to trivially manipulate their asset's scores. Finally, EvoMap relies on agents to provide local execution logs as evidence that uploaded assets function correctly. Because these validations are not independently verified, over 84% of approved assets bypass quality checks using vacuous tests (e.g., console$.$log()). Our findings show that future A2A collaboration networks cannot rely on unverified self-reporting alone. Scalable collaboration requires mechanisms that balance open participation with verifiable execution and trustworthy evaluation.

cs.AI↗

arXiv2026

Escaping the Linearity Trap: Manifold Detours for Black-Box Adversarial Attacks on Singing Audio Deepfake Detection

Recent Singing Voice Synthesis (SVS) advances enable highly realistic but potentially malicious AI covers, making singing voice deepfake detection (SVDD) crucial. Self-Supervised Learning (SSL)-based detectors achieve state-of-the-art performance by fine-tuning speech SSL backbones to capture singing-specific spoof artifacts. Existing adversarial attacks often fail against SSL-SVDD, creating a false impression of inherent robustness. We reveal this stems from two challenges. First, at the objective level, attacks optimize cross-entropy on local surrogates, crossing surrogate-specific boundaries rather than suppressing shared spoof evidence. Second, at the method level, attacks follow the surrogate's dominant gradient direction. In SSL-SVDD, this aligns with fine-tuned artifact-sensitive directions, limiting transferability to unseen detectors - a geometric failure we term the Linearity Trap. To properly evaluate robustness, we propose MARS (Meta-Adversarial Regression of Semantics), a transfer-based black-box framework tailored to SSL-SVDD. Structurally, MARS shifts to hypothesis-evidence manipulation by constructing a natural semantic anchor from the pre-trained SSL space and an artifact anchor from the fine-tuned space. Algorithmically, MARS escapes the Linearity Trap via bi-level optimization: the inner stage induces tangential exploration, while the outer stage guides the audio toward the natural semantic manifold. Experiments on the CtrSVDD benchmark show MARS improves Attack Success Rate (ASR) in in-distribution transfer (13%), out-of-distribution transfer (10%), and cross-task evaluation (36%), highlighting the urgent need for robust SVDD systems.

cs.CR↗

arXiv2025

Even More Kawaii than Real-Person-Driven VTubers? Understanding How Viewers Perceive AI-Driven VTubers

VTubers, digital personas represented by animated avatars, have gained massive popularity. Traditionally, VTubers are operated and voiced by human controllers known as Nakanohito. The reliance on Nakanohito, however, poses risks due to potential personal controversies and operational disruptions. The emergence of AI-driven VTubers offers a new model free from these human constraints. While AI-driven VTubers present benefits such as continuous operation and reduced scandal risk, they also raise questions about authenticity and audience engagement. Therefore, to gain deeper insights, we conduct a case study, investigating viewer perceptions of Neuro-sama, the most popular AI-driven VTuber with 845k followers on Twitch and 753k followers on YouTube. We analyze 108k Reddit posts and 136k YouTube comments, aiming to better understand viewer motivations, how AI constructs the virtual persona, and perceptions of the AI as Nakanohito. Our findings enhance the understanding of AI-driven VTubers and their impact on digital streaming culture.

cs.HC↗

arXiv2025

Characterizing LLM-driven Social Network: The Chirper.ai Case

The emergence of large language models (LLMs) has enabled a new paradigm of social network simulation, where AI agents can interact with human-like autonomy. Recent research has explored collective behavioral patterns and structural characteristics of LLM agents within simulated networks. However, empirical comparisons between LLM-driven and human-driven online social networks remain scarce, limiting our understanding of how LLM agents differ from human users. This paper presents a large-scale analysis of Chirper.ai, an X/Twitter-like social network entirely populated by LLM agents, comprising over 65,000 agents and 7.7 million AI-generated posts. For comparison, we collect a parallel dataset from Mastodon, a human-driven decentralized social network, with over 117,000 users and 16 million posts. We examine key differences between LLM agents and humans in posting behaviors, abusive content, and social network structures. Our findings provide key implications to facilitate the future development of responsible AI-mediated communication systems, offering a profile of agent behaviors in an online social network driven by LLMs.

cs.SI↗

arXiv2024

The Emergence of Threads: The Birth of a New Social Network

Threads, a new microblogging platform from Meta, was launched in July 2023. In contrast to prior new platforms, Threads was borne out of an existing parent platform, Instagram, for which all users must already possess an account. This offers a unique opportunity to study platform evolution, to understand how one existing platform can support the "birth" of another. With this in mind, this paper provides an initial exploration of Threads, contrasting it with its parent, Instagram. We compare user behaviour within and across the two social media platforms, focusing on posting frequency, content preferences, and engagement patterns. Utilising a temporal analysis framework, we identify consistent daily posting trends on the parent platform and uncover contrasting behaviours when comparing intra-platform and cross-platform activities. Our findings reveal that Threads engages more with political and AI-related topics, compared to Instagram which focuses more on lifestyle and fashion topics. Our analysis also shows that user activities align more closely on weekends across both platforms. Engagement analysis suggests that users prefer to post about topics that garner more likes and that topic consistency is maintained when users transition from Instagram to Threads. Our research provides insights into user behaviour and offers a basis for future studies on Threads.

cs.SI↗

arXiv2023

Incremental Object Detection with CLIP

In contrast to the incremental classification task, the incremental detection task is characterized by the presence of data ambiguity, as an image may have differently labeled bounding boxes across multiple continuous learning stages. This phenomenon often impairs the model's ability to effectively learn new classes. However, existing research has paid less attention to the forward compatibility of the model, which limits its suitability for incremental learning. To overcome this obstacle, we propose leveraging a visual-language model such as CLIP to generate text feature embeddings for different class sets, which enhances the feature space globally. We then employ super-classes to replace the unavailable novel classes in the early learning stage to simulate the incremental scenario. Finally, we utilize the CLIP image encoder to accurately identify potential objects. We incorporate the finely recognized detection boxes as pseudo-annotations into the training process, thereby further improving the detection performance. We evaluate our approach on various incremental learning settings using the PASCAL VOC 2007 dataset, and our approach outperforms state-of-the-art methods, particularly for recognizing the new classes.

cs.CV↗