arXiv subjects

Bowei Chen

Publications and source records attributed to Bowei Chen.

At least 19 recordsLinked to original sources

arXiv2026

Layering Virtual Try-On

In the real world, fashion is about layering: adding a jacket over a shirt, or a sequence of adding and removing layers, rather than just a single-layer swap. This fundamental real-world task remains a challenge in existing Virtual Try-On (VTON) methods, which excel at single-layer replacement but are not designed to layer or de-layer an existing outfit. This paper proposes Layering Virtual Try-On (LVTON), a layering benchmark and method that preserves an existing outfit while enabling sequential layering. We find that current VTON paradigms are fundamentally ill-equipped for LVTON, as their reliance on cloth-agnostic representations and single-item datasets discards essential layering context. Our key insight is that the LVTON challenge must be disentangled into two distinct competencies: (1) General VTON Priors (e.g., deformation, identity preservation) and (2) Specific Layering Knowledge (e.g., layering order and occlusion reasoning). First, our model obtains general VTON priors by being trained on data produced by an automatic data generation pipeline that synthesizes samples from fashion videos via segmentation and inpainting. Second, the model is fine-tuned on a small, dedicated LVTON dataset to learn the layering logic. Our method achieves state-of-the-art results on our LVTON benchmark and demonstrates superior generalizability on traditional VTON benchmarks, setting new state-of-the-art results when fine-tuned and exhibiting zero-shot capabilities.

cs.CV↗

arXiv2026

DeXposure-Claw: An Agentic System for DeFi Risk Supervision

Decentralized finance exposes supervisors to fast-moving, networked credit risks. General-purpose LLM agents fit this setting poorly: they over-read weak evidence and recommend high-stakes interventions, while existing evaluations offer no regulator-aligned way to measure the resulting false alarms. We introduce DeXposure-Claw, a forecast-grounded agentic supervision system that routes LLM decisions through structured evidence: (1) DeXposure-FM, a graph time-series foundation model, forecasts future exposure networks; (2) deterministic monitors and stress scenarios then turn those forecasts into typed alerts, attribution signals, and scenario evidence; and (3) data-health and confidence gates constrain escalation before DeXposure-Claw emits auditable supervisory tickets with rationales. We further develop DeXposure-Bench, a six-axis evaluation harness, whose decision axis scores tickets against a regulator-aligned absolute-loss ground truth and an explicit false-intervention rate. Experiments on five years of weekly real data fully support our system. Code is at https://github.com/EVIEHub/DeXposure-Claw.

cs.AI↗

arXiv2025

Cortex AISQL: A Production SQL Engine for Unstructured Data

Snowflake's Cortex AISQL is a production SQL engine that integrates native semantic operations directly into SQL. This integration allows users to write declarative queries that combine relational operations with semantic reasoning, enabling them to query both structured and unstructured data effortlessly. However, making semantic operations efficient at production scale poses fundamental challenges. Semantic operations are more expensive than traditional SQL operations, possess distinct latency and throughput characteristics, and their cost and selectivity are unknown during query compilation. Furthermore, existing query engines are not designed to optimize semantic operations. The AISQL query execution engine addresses these challenges through three novel techniques informed by production deployment data from Snowflake customers. First, AI-aware query optimization treats AI inference cost as a first-class optimization objective, reasoning about large language model (LLM) cost directly during query planning to achieve 2-8$\times$ speedups. Second, adaptive model cascades reduce inference costs by routing most rows through a fast proxy model while escalating uncertain cases to a powerful oracle model, achieving 2-6$\times$ speedups while maintaining 90-95% of oracle model quality. Third, semantic join query rewriting lowers the quadratic time complexity of join operations to linear through reformulation as multi-label classification tasks, achieving 15-70$\times$ speedups with often improved prediction quality. AISQL is deployed in production at Snowflake, where it powers diverse customer workloads across analytics, search, and content understanding.

cs.DB↗

arXiv2025

AlignTok: Aligning Visual Foundation Encoders to Tokenizers for Diffusion Models

In this work, we propose aligning pretrained visual encoders to serve as tokenizers for latent diffusion models in image generation. Unlike training a variational autoencoder (VAE) from scratch, which primarily emphasizes low-level details, our approach leverages the rich semantic structure of foundation encoders. We introduce a three-stage alignment strategy called AlignTok: (1) freeze the encoder and train an adapter and a decoder to establish a semantic latent space; (2) jointly optimize all components with an additional semantic preservation loss, enabling the encoder to capture perceptual details while retaining high-level semantics; and (3) refine the decoder for improved reconstruction quality. This alignment yields semantically rich image tokenizers that benefit diffusion models. On ImageNet 256$\times$256, our tokenizer accelerates the convergence of diffusion models, reaching a gFID of 1.90 within just 64 epochs, and improves generation both with and without classifier-free guidance. Scaling to LAION, text-to-image models trained with our tokenizer consistently outperforms FLUX VAE and VA-VAE under the same training steps. Overall, our method is simple, scalable, and establishes a semantically grounded paradigm for continuous tokenizer design.

cs.CV↗

arXiv2025

Generating Fit Check Videos with a Handheld Camera

Self-captured full-body videos are popular, but most deployments require mounted cameras, carefully-framed shots, and repeated practice. We propose a more convenient solution that enables full-body video capture using handheld mobile devices. Our approach takes as input two static photos (front and back) of you in a mirror, along with an IMU motion reference that you perform while holding your mobile phone, and synthesizes a realistic video of you performing a similar target motion. We enable rendering into a new scene, with consistent illumination and shadows. We propose a novel video diffusion-based model to achieve this. Specifically, we propose a parameter-free frame generation strategy and a multi-reference attention mechanism to effectively integrate appearance information from both the front and back selfies into the video diffusion model. Further, we introduce an image-based fine-tuning strategy to enhance frame sharpness and improve shadows and reflections generation for more realistic human-scene composition.

cs.CV↗

arXiv2024

Inverse Painting: Reconstructing The Painting Process

Given an input painting, we reconstruct a time-lapse video of how it may have been painted. We formulate this as an autoregressive image generation problem, in which an initially blank "canvas" is iteratively updated. The model learns from real artists by training on many painting videos. Our approach incorporates text and region understanding to define a set of painting "instructions" and updates the canvas with a novel diffusion-based renderer. The method extrapolates beyond the limited, acrylic style paintings on which it has been trained, showing plausible results for a wide range of artistic styles and genres.

cs.CV↗

arXiv2024

Learning Feature-Preserving Portrait Editing from Generated Pairs

Portrait editing is challenging for existing techniques due to difficulties in preserving subject features like identity. In this paper, we propose a training-based method leveraging auto-generated paired data to learn desired editing while ensuring the preservation of unchanged subject features. Specifically, we design a data generation process to create reasonably good training pairs for desired editing at low cost. Based on these pairs, we introduce a Multi-Conditioned Diffusion Model to effectively learn the editing direction and preserve subject features. During inference, our model produces accurate editing mask that can guide the inference process to further preserve detailed subject features. Experiments on costume editing and cartoon expression editing show that our method achieves state-of-the-art quality, quantitatively and qualitatively.

cs.CV↗

arXiv2024

EcoVal: An Efficient Data Valuation Framework for Machine Learning

Quantifying the value of data within a machine learning workflow can play a pivotal role in making more strategic decisions in machine learning initiatives. The existing Shapley value based frameworks for data valuation in machine learning are computationally expensive as they require considerable amount of repeated training of the model to obtain the Shapley value. In this paper, we introduce an efficient data valuation framework EcoVal, to estimate the value of data for machine learning models in a fast and practical manner. Instead of directly working with individual data sample, we determine the value of a cluster of similar data points. This value is further propagated amongst all the member cluster points. We show that the overall value of the data can be determined by estimating the intrinsic and extrinsic value of each data. This is enabled by formulating the performance of a model as a \textit{production function}, a concept which is popularly used to estimate the amount of output based on factors like labor and capital in a traditional free economic market. We provide a formal proof of our valuation technique and elucidate the principles and mechanisms that enable its accelerated performance. We demonstrate the real-world applicability of our method by showcasing its effectiveness for both in-distribution and out-of-sample data. This work addresses one of the core challenges of efficient data valuation at scale in machine learning models. The code is available at \underline{https://github.com/respai-lab/ecoval}.

cs.LG↗

arXiv2023

SRNI-CAR: A comprehensive dataset for analyzing the Chinese automotive market

The automotive industry plays a critical role in the global economy, and particularly important is the expanding Chinese automobile market due to its immense scale and influence. However, existing automotive sector datasets are limited in their coverage, failing to adequately consider the growing demand for more and diverse variables. This paper aims to bridge this data gap by introducing a comprehensive dataset spanning the years from 2016 to 2022, encompassing sales data, online reviews, and a wealth of information related to the Chinese automotive industry. This dataset serves as a valuable resource, significantly expanding the available data. Its impact extends to various dimensions, including improving forecasting accuracy, expanding the scope of business applications, informing policy development and regulation, and advancing academic research within the automotive sector. To illustrate the dataset's potential applications in both business and academic contexts, we present two application examples. Our developed dataset enhances our understanding of the Chinese automotive market and offers a valuable tool for researchers, policymakers, and industry stakeholders worldwide.

econ.GN↗

arXiv2023

Total Selfie: Generating Full-Body Selfies

We present a method to generate full-body selfies from photographs originally taken at arms length. Because self-captured photos are typically taken close up, they have limited field of view and exaggerated perspective that distorts facial shapes. We instead seek to generate the photo some one else would take of you from a few feet away. Our approach takes as input four selfies of your face and body, a background image, and generates a full-body selfie in a desired target pose. We introduce a novel diffusion-based approach to combine all of this information into high-quality, well-composed photos of you with the desired pose and background.

cs.CV↗

arXiv2022

Learning Continuous Implicit Representation for Near-Periodic Patterns

Near-Periodic Patterns (NPP) are ubiquitous in man-made scenes and are composed of tiled motifs with appearance differences caused by lighting, defects, or design elements. A good NPP representation is useful for many applications including image completion, segmentation, and geometric remapping. But representing NPP is challenging because it needs to maintain global consistency (tiled motifs layout) while preserving local variations (appearance differences). Methods trained on general scenes using a large dataset or single-image optimization struggle to satisfy these constraints, while methods that explicitly model periodicity are not robust to periodicity detection errors. To address these challenges, we learn a neural implicit representation using a coordinate-based MLP with single image optimization. We design an input feature warping module and a periodicity-guided patch loss to handle both global consistency and local variations. To further improve the robustness, we introduce a periodicity proposal module to search and use multiple candidate periodicities in our pipeline. We demonstrate the effectiveness of our method on more than 500 images of building facades, friezes, wallpapers, ground, and Mondrian patterns on single and multi-planar scenes.

cs.CV↗

arXiv2022

Semantically Supervised Appearance Decomposition for Virtual Staging from a Single Panorama

We describe a novel approach to decompose a single panorama of an empty indoor environment into four appearance components: specular, direct sunlight, diffuse and diffuse ambient without direct sunlight. Our system is weakly supervised by automatically generated semantic maps (with floor, wall, ceiling, lamp, window and door labels) that have shown success on perspective views and are trained for panoramas using transfer learning without any further annotations. A GAN-based approach supervised by coarse information obtained from the semantic map extracts specular reflection and direct sunlight regions on the floor and walls. These lighting effects are removed via a similar GAN-based approach and a semantic-aware inpainting step. The appearance decomposition enables multiple applications including sun direction estimation, virtual furniture insertion, floor material replacement, and sun direction change, providing an effective tool for virtual home staging. We demonstrate the effectiveness of our approach on a large and recently released dataset of panoramas of empty homes.

cs.GR↗

arXiv2021

DVM-CAR: A large-scale automotive dataset for visual marketing research and applications

There is a growing interest in product aesthetics analytics and design. However, the lack of available large-scale data that covers various variables and information is one of the biggest challenges faced by analysts and researchers. In this paper, we present our multidisciplinary initiative of developing a comprehensive automotive dataset from different online sources and formats. Specifically, the created dataset contains 1.4 million images from 899 car models and their corresponding model specifications and sales information over more than ten years in the UK market. Our work makes significant contributions to: (i) research and applications in the automotive industry; (ii) big data creation and sharing; (iii) database design; and (iv) data fusion. Apart from our motivation, technical details and data structure, we further present three simple examples to demonstrate how our data can be used in business research and applications.

cs.CV↗

arXiv2021

Learning Robust Variational Information Bottleneck with Reference

We propose a new approach to train a variational information bottleneck (VIB) that improves its robustness to adversarial perturbations. Unlike the traditional methods where the hard labels are usually used for the classification task, we refine the categorical class information in the training phase with soft labels which are obtained from a pre-trained reference neural network and can reflect the likelihood of the original class labels. We also relax the Gaussian posterior assumption in the VIB implementation by using the mutual information neural estimation. Extensive experiments have been performed with the MNIST and CIFAR-10 datasets, and the results show that our proposed approach significantly outperforms the benchmarked models.

cs.LG↗

arXiv2020

Multi-Task Variational Information Bottleneck

Multi-task learning (MTL) is an important subject in machine learning and artificial intelligence. Its applications to computer vision, signal processing, and speech recognition are ubiquitous. Although this subject has attracted considerable attention recently, the performance and robustness of the existing models to different tasks have not been well balanced. This article proposes an MTL model based on the architecture of the variational information bottleneck (VIB), which can provide a more effective latent representation of the input features for the downstream tasks. Extensive observations on three public data sets under adversarial attacks show that the proposed model is competitive to the state-of-the-art algorithms concerning the prediction accuracy. Experimental results suggest that combining the VIB and the task-dependent uncertainties is a very effective way to abstract valid information from the input features for accomplishing multiple tasks.

cs.LG↗

arXiv2019

A hybrid model for predicting human physical activity status from lifelogging data

One trend in the recent healthcare transformations is people are encouraged to monitor and manage their health based on their daily diets and physical activity habits. However, much attention of the use of operational research and analytical models in healthcare has been paid to the systematic level such as country or regional policy making or organisational issues. This paper proposes a model concerned with healthcare analytics at the individual level, which can predict human physical activity status from sequential lifelogging data collected from wearable sensors. The model has a two-stage hybrid structure (in short, MOGP-HMM) -- a multi-objective genetic programming (MOGP) algorithm in the first stage to reduce the dimensions of lifelogging data and a hidden Markov model (HMM) in the second stage for activity status prediction over time. It can be used as a decision support tool to provide real-time monitoring, statistical analysis and personalized advice to individuals, encouraging positive attitudes towards healthy lifestyles. We validate the model with the real data collected from a group of participants in the UK, and compare it with other popular two-stage hybrid models. Our experimental results show that the MOGP-HMM can achieve comparable performance. To the best of our knowledge, this is the very first study that uses the MOGP in the hybrid two-stage structure for individuals' activity status prediction. It fits seamlessly with the current trend in the UK healthcare transformation of patient empowerment as well as contributing to a strategic development for more efficient and cost-effective provision of healthcare.

cs.CY↗

arXiv2019

Combining guaranteed and spot markets in display advertising: Selling guaranteed page views with stochastic demand

While page views are often sold instantly through real-time auctions when users visit websites, they can also be sold in advance via guaranteed contracts. In this paper, we present a dynamic programming model to study how an online publisher should optimally allocate and price page views between guaranteed and spot markets. The problem is challenging because the allocation and pricing of guaranteed contracts affect how advertisers split their purchases between the two markets, and the terminal value of the model is endogenously determined by the updated dual force of supply and demand in auctions. We take the advertisers' purchasing behaviour into consideration, i.e., risk aversion and stochastic demand arrivals, and present a scalable and efficient algorithm for the optimal solution. The model is also empirically validated with a commercial dataset. The experimental results show that selling page views via both channels can increase the publisher's expected total revenue, and the optimal pricing and allocation strategies are robust to different market and advertiser types.

cs.GT↗

arXiv2019

Index Tracking with Cardinality Constraints: A Stochastic Neural Networks Approach

Partial (replication) index tracking is a popular passive investment strategy. It aims to replicate the performance of a given index by constructing a tracking portfolio which contains some constituents of the index. The tracking error optimisation is quadratic and NP-hard when taking the L0 constraint into account so it is usually solved by heuristic methods such as evolutionary algorithms. This paper introduces a simple, efficient and scalable connectionist model as an alternative. We propose a novel reparametrisation method and then solve the optimisation problem with stochastic neural networks. The proposed approach is examined with S&P 500 index data for more than 10 years and compared with widely used index tracking approaches such as forward and backward selection and the largest market capitalisation methods. The empirical results show our model achieves excellent performance. Compared with the benchmarked models, our model has the lowest tracking error, across a range of portfolio sizes. Meanwhile it offers comparable performance to the others on secondary criteria such as volatility, Sharpe ratio and maximum drawdown.

q-fin.PM↗