arXiv subjects

Xi Yin

Publications and source records attributed to Xi Yin.

At least 19 recordsLinked to original sources

arXiv2026

Moving Alphabet: A Controlled Study of Training Data for Text-to-Video Generation

Text-to-video generation has advanced significantly over the past five years through scaling of model size, data, and compute. Unlike model architecture, training data is often underexplored. Real-world data curation is complex and non-trivial, involving clip selection from raw videos and captioning to create video-text pairs for learning text-to-video mappings. We study how data distribution and caption quality impact text-to-video models. To enable controlled experiments, we introduce Moving Alphabet, a procedural testbed that renders letters with varying fonts, colors, sizes, and positions, moving in different directions and speeds against a black background. This design allows precise control over data distribution and caption quality by corrupting ground-truth metadata. Our experiments yield three findings: a) a diverse and balanced distribution of video content and duration is critical for generalization; b) caption quality significantly affects both model performance and training efficiency, suggesting that text-to-video models are bounded by video understanding capabilities; and c) classifier-free guidance and fine-tuning on high-quality data provide partial recovery from models trained on corrupted captions, but cannot fully compensate for poor pre-training data. We believe these insights can inform the development of large-scale text-to-video models, and we advocate for greater attention to the science of pre-training data.

cs.CV↗

arXiv2026

Closed String Field Theory in 25.99 Dimensions

We return to and refine Zwiebach's formulation of closed string field theory (CSFT) built around non-critical backgrounds [1,2], restricting our attention to genus zero. The structure involves a special string state $F$ that encodes the failure of worldsheet BRST invariance, and a metric-dependent descent operator $\mathcal{B}$ adapted to the Weyl frame. We construct the mixed moduli spaces needed for the classical BV action, prove their existence, and extend the Sen-Zwiebach background independence argument to first order off of the conformal locus. We apply the formalism to the mildest deviation away from criticality - worldsheet CFTs with nonzero central charge: we consider both D=26-$\epsilon$ dimensional flat space and linear dilaton profiles in bosonic string theory, focusing for simplicity on building solutions that depend on only one of the D dimensions.

hep-th↗

arXiv2026

D-instanton Effects on a D3-brane

It has been proposed by Sen that the D-instanton effects in string theory can be systematically determined through the framework of open-closed string field theory. We apply the latter formalism to analyze the D-instanton corrections to the quantum effective action of a D3-brane in type IIB superstring theory, and determine the leading single and multi-instanton contributions to the $D^4 F^4$ effective coupling which is unprotected by supersymmetry. Notably, while we find that the one-instanton contribution agrees with a conjecture of Green and Gutperle, the multi-instanton contribution disagrees with the conjecture.

hep-th↗

arXiv2026

Human detectors are surprisingly powerful reward models

Video generation models have recently achieved impressive visual fidelity and temporal coherence. Yet, they continue to struggle with complex, non-rigid motions, especially when synthesizing humans performing dynamic actions such as sports, dance, etc. Generated videos often exhibit missing or extra limbs, distorted poses, or physically implausible actions. In this work, we propose a remarkably simple reward model, HuDA, to quantify and improve the human motion in generated videos. HuDA integrates human detection confidence for appearance quality, and a temporal prompt alignment score to capture motion realism. We show this simple reward function that leverages off-the-shelf models without any additional training, outperforms specialized models finetuned with manually annotated data. Using HuDA for Group Reward Policy Optimization (GRPO) post-training of video models, we significantly enhance video generation, especially when generating complex human motions, outperforming state-of-the-art models like Wan 2.1, with win-rate of 73%. Finally, we demonstrate that HuDA improves generation quality beyond just humans, for instance, significantly improving generation of animal videos and human-object interactions.

cs.CV↗

arXiv2026

On the Flux Sectors of Matrix String Theory

We analyze the gapped flux vacua of 2D (8,8) SU(N) super-Yang-Mills theory. Based on the matrix string theory duality, we conjecture the spectrum of massive resonances in the gauge theory at large N and beyond the 't Hooft scaling regime.

hep-th↗

arXiv2025

On the $AdS_5\times S^5$ Solution of Superstring Field Theory

We determine the $AdS_5\times S^5$ solution of type IIB superstring field theory (SFT) to the third order in the expansion with respect to Ramond-Ramond (RR) flux, demonstrate its supersymmetry from the SFT gauge transformations, and identify the massless RR axion in the spectrum of linearized fluctuations. We present an all-order solution in the pp-wave limit and comment on potential obstructions to, and the existence of, the all-order $AdS_5\times S^5$ solution.

hep-th↗

arXiv2025

BlueLM-2.5-3B Technical Report

We present BlueLM-2.5-3B, a compact and unified dense Multimodal Large Language Model (MLLM) designed for efficient edge-device deployment, offering strong general-purpose and reasoning capabilities. To the best of our knowledge, this is the first 3B-scale MLLM to support both thinking and non-thinking modes, while also enabling explicit control over thinking token budget. BlueLM-2.5-3B is developed through diversified data curation, key data resampling, hybrid heterogeneous reinforcement learning, and a high-performance training infrastructure. Our model achieves superior multimodal capacity while preserving competitive pure-text performance with only 2.9 billion parameters. We conduct comprehensive evaluations across a broad range of multimodal and text-only benchmarks. In thinking mode, BlueLM-2.5-3B achieves comparable performance to Qwen3-4B on text-only benchmarks, and trails the larger Kimi-VL-A3B-16B by only about 5% on average across multimodal evaluations. In non-thinking mode, it outperforms Qwen2.5-VL-3B on the majority of multimodal benchmarks. Additionally, BlueLM-2.5-3B exhibits exceptional data efficiency. All of the aforementioned performance is achieved with substantially less total training data than Qwen2.5-VL-3B and Qwen3-4B. We hope our work contributes to the advancement of high-performance, on-device MLLMs and provides meaningful insights to the research community.

cs.AI↗

arXiv2025

Diffeomorphism in Closed String Field Theory

We explain how the spacetime diffeomorphism in the classical bosonic closed string field theory (SFT) is represented as $L_\infty$ gauge transformations in weakly curved backgrounds. In particular, we demonstrate the explicit map between covariant spacetime fields and the string fields in the flat-vertex frame to the leading few orders in perturbation theory. We further consider a Batalin-Vilkovisky system of extended string fields that manifests diffeomorphism invariance and allows for extending certain perturbative SFT solutions to the large field regime.

hep-th↗

arXiv2025

UI-R1: Enhancing Efficient Action Prediction of GUI Agents by Reinforcement Learning

The recent DeepSeek-R1 has showcased the emergence of reasoning capabilities in LLMs through reinforcement learning (RL) with rule-based rewards. Despite its success in language models, its application in multi-modal domains, particularly in graphic user interface (GUI) agent tasks, remains under-explored. To address this issue, we propose UI-R1, the first framework to explore how rule-based RL can enhance the reasoning capabilities of multimodal large language models (MLLMs) for GUI action prediction tasks. Specifically, UI-R1 introduces a novel rule-based action reward, enabling model optimization via policy-based algorithms such as Group Relative Policy Optimization (GRPO). For efficient training, we curate a small yet high-quality dataset of 136 challenging tasks, encompassing five common action types on mobile devices. Experimental results demonstrate that our proposed UI-R1-3B achieves significant improvements over the base model (i.e. Qwen2.5-VL-3B) on both in-domain (ID) and out-of-domain (OOD) tasks, with average accuracy gains of 22.1% on ScreenSpot, 6.0% on ScreenSpot-Pro, and 12.7% on ANDROIDCONTROL. Furthermore, UI-R1-3B delivers competitive performance compared to larger models (e.g., OS-Atlas-7B) trained via supervised fine-tuning (SFT) on 76K samples. We additionally develop an optimized version, UI-R1-E-3B, which significantly improves both grounding efficiency and accuracy. These results underscore the potential of rule-based reinforcement learning to advance GUI understanding and control, paving the way for future research in this domain. Code website: https://github.com/lll6gg/UI-R1.

cs.AI↗

arXiv2025

Generating Multi-Image Synthetic Data for Text-to-Image Customization

Customization of text-to-image models enables users to insert new concepts or objects and generate them in unseen settings. Existing methods either rely on comparatively expensive test-time optimization or train encoders on single-image datasets without multi-image supervision, which can limit image quality. We propose a simple approach to address these challenges. We first leverage existing text-to-image models and 3D datasets to create a high-quality Synthetic Customization Dataset (SynCD) consisting of multiple images of the same object in different lighting, backgrounds, and poses. Using this dataset, we train an encoder-based model that incorporates fine-grained visual details from reference images via a shared attention mechanism. Finally, we propose an inference technique that normalizes text and image guidance vectors to mitigate overexposure issues in sampled images. Through extensive experiments, we show that our encoder-based model, trained on SynCD, and with the proposed inference algorithm, improves upon existing encoder-based methods on standard customization benchmarks.

cs.CV↗

arXiv2024

MotiF: Making Text Count in Image Animation with Motion Focal Loss

Text-Image-to-Video (TI2V) generation aims to generate a video from an image following a text description, which is also referred to as text-guided image animation. Most existing methods struggle to generate videos that align well with the text prompts, particularly when motion is specified. To overcome this limitation, we introduce MotiF, a simple yet effective approach that directs the model's learning to the regions with more motion, thereby improving the text alignment and motion generation. We use optical flow to generate a motion heatmap and weight the loss according to the intensity of the motion. This modified objective leads to noticeable improvements and complements existing methods that utilize motion priors as model inputs. Additionally, due to the lack of a diverse benchmark for evaluating TI2V generation, we propose TI2V Bench, a dataset consists of 320 image-text pairs for robust evaluation. We present a human evaluation protocol that asks the annotators to select an overall preference between two videos followed by their justifications. Through a comprehensive evaluation on TI2V Bench, MotiF outperforms nine open-sourced models, achieving an average preference of 72%. The TI2V Bench and additional results are released in https://wang-sj16.github.io/motif/.

cs.CV↗

arXiv2024

Flowing from Words to Pixels: A Noise-Free Framework for Cross-Modality Evolution

Diffusion models, and their generalization, flow matching, have had a remarkable impact on the field of media generation. Here, the conventional approach is to learn the complex mapping from a simple source distribution of Gaussian noise to the target media distribution. For cross-modal tasks such as text-to-image generation, this same mapping from noise to image is learnt whilst including a conditioning mechanism in the model. One key and thus far relatively unexplored feature of flow matching is that, unlike Diffusion models, they are not constrained for the source distribution to be noise. Hence, in this paper, we propose a paradigm shift, and ask the question of whether we can instead train flow matching models to learn a direct mapping from the distribution of one modality to the distribution of another, thus obviating the need for both the noise distribution and conditioning mechanism. We present a general and simple framework, CrossFlow, for cross-modal flow matching. We show the importance of applying Variational Encoders to the input data, and introduce a method to enable Classifier-free guidance. Surprisingly, for text-to-image, CrossFlow with a vanilla transformer without cross attention slightly outperforms standard flow matching, and we show that it scales better with training steps and model size, while also allowing for interesting latent arithmetic which results in semantically meaningful edits in the output space. To demonstrate the generalizability of our approach, we also show that CrossFlow is on par with or outperforms the state-of-the-art for various cross-modal / intra-modal mapping tasks, viz. image captioning, depth estimation, and image super-resolution. We hope this paper contributes to accelerating progress in cross-modal media generation.

cs.CV↗

arXiv2024

Movie Gen: A Cast of Media Foundation Models

We present Movie Gen, a cast of foundation models that generates high-quality, 1080p HD videos with different aspect ratios and synchronized audio. We also show additional capabilities such as precise instruction-based video editing and generation of personalized videos based on a user's image. Our models set a new state-of-the-art on multiple tasks: text-to-video synthesis, video personalization, video editing, video-to-audio generation, and text-to-audio generation. Our largest video generation model is a 30B parameter transformer trained with a maximum context length of 73K video tokens, corresponding to a generated video of 16 seconds at 16 frames-per-second. We show multiple technical innovations and simplifications on the architecture, latent spaces, training objectives and recipes, data curation, evaluation protocols, parallelization techniques, and inference optimizations that allow us to reap the benefits of scaling pre-training data, model size, and training compute for training large scale media generation models. We hope this paper helps the research community to accelerate progress and innovation in media generation models. All videos from this paper are available at https://go.fb.me/MovieGenResearchVideos.

cs.CV↗

arXiv2024

Thermal Bootstrap of Matrix Quantum Mechanics

We implement a bootstrap method that combines stationary state conditions, thermal inequalities, and semidefinite relaxations of matrix logarithm in the ungauged one-matrix quantum mechanics, at finite rank N as well as in the large N limit, and determine finite temperature observables that interpolate between available analytic results in the low and high temperature limits respectively. We also obtain bootstrap bounds on thermal phase transition as well as preliminary results in the ungauged two-matrix quantum mechanics.

hep-th↗

arXiv2024

Proactive Schemes: A Survey of Adversarial Attacks for Social Good

Adversarial attacks in computer vision exploit the vulnerabilities of machine learning models by introducing subtle perturbations to input data, often leading to incorrect predictions or classifications. These attacks have evolved in sophistication with the advent of deep learning, presenting significant challenges in critical applications, which can be harmful for society. However, there is also a rich line of research from a transformative perspective that leverages adversarial techniques for social good. Specifically, we examine the rise of proactive schemes-methods that encrypt input data using additional signals termed templates, to enhance the performance of deep learning models. By embedding these imperceptible templates into digital media, proactive schemes are applied across various applications, from simple image enhancements to complicated deep learning frameworks to aid performance, as compared to the passive schemes, which don't change the input data distribution for their framework. The survey delves into the methodologies behind these proactive schemes, the encryption and learning processes, and their application to modern computer vision and natural language processing applications. Additionally, it discusses the challenges, potential vulnerabilities, and future directions for proactive schemes, ultimately highlighting their potential to foster the responsible and secure advancement of deep learning technologies.

cs.CV↗

arXiv2024

Conformal Perturbation Theory and Tachyon-Dilaton Eschatology via String Fields

We analyze deformations of two-dimensional conformal field theory (CFT) from the perspective of classical bosonic closed string field theory (SFT). The latter can be viewed as a version of Wilsonian renormalization group (RG) improved conformal perturbation theory, where the renormalization scheme is defined through the choice of string vertices in the construction of SFT. Furthermore, the CFT data at the RG fixed point can be recovered from the spectrum and amplitudes of string field fluctuations. As applications, we construct the Horowitz-Polchinski "string star" solution in SFT, and a solution of tachyon-dilaton condensation that deforms the noncompact boson to minimal models by creating a pair of "Runkel-Watts walls".

hep-th↗

arXiv2023

Bootstrapping M-theory Orbifolds

We analyze correlation functions of $SU(k) \times SU(2)_F$ flavor currents in a family of three-dimensional ${\cal N}=4$ superconformal field theories, combining analytic bootstrap methods with input from supersymmetric localization. Via holographic duality, we extract gluon and graviton scattering amplitudes of M-theory on ${\rm AdS}_4\times S^7/\mathbb{Z}_k$ which contains a $\mathbb{C}^2/\mathbb{Z}_{k}$ orbifold singularity. From these results, we derive aspects of the effective description of M-theory on the orbifold singularity beyond its leading low energy limit. We also determine a threshold correction to the holographic correlator, which vanishes due to equal and opposite contributions from the two-loop gluon and the tree-level bulk graviton exchange.

hep-th↗

arXiv2023

AcademicGPT: Empowering Academic Research

Large Language Models (LLMs) have demonstrated exceptional capabilities across various natural language processing tasks. Yet, many of these advanced LLMs are tailored for broad, general-purpose applications. In this technical report, we introduce AcademicGPT, designed specifically to empower academic research. AcademicGPT is a continual training model derived from LLaMA2-70B. Our training corpus mainly consists of academic papers, thesis, content from some academic domain, high-quality Chinese data and others. While it may not be extensive in data scale, AcademicGPT marks our initial venture into a domain-specific GPT tailored for research area. We evaluate AcademicGPT on several established public benchmarks such as MMLU and CEval, as well as on some specialized academic benchmarks like PubMedQA, SCIEval, and our newly-created ComputerScienceQA, to demonstrate its ability from general knowledge ability, to Chinese ability, and to academic ability. Building upon AcademicGPT's foundation model, we also developed several applications catered to the academic area, including General Academic Question Answering, AI-assisted Paper Reading, Paper Review, and AI-assisted Title and Abstract Generation.

cs.CL↗