Wenyi Wu

Sibo Zhu, Wenyi Wu, Kun Zhou et al.

The remarkable progress of vision-language models (VLMs) has enabled GUI agents to interact with computers in a human-like manner. Yet real-world computer-use tasks remain difficult due to long-horizon workflows, diverse interfaces, and frequent intermediate errors. Prior work equips agents with external memory built from large collections of trajectories, but relies on flat retrieval over discrete summaries or continuous embeddings, falling short of the structured organization and self-evolving characteristics of human memory. Inspired by the brain, we propose Hybrid Self-evolving Structured Memory (HyMEM), a graph-based memory that couples discrete high-level symbolic nodes with continuous trajectory embeddings. HyMEM maintains a graph structure to support multi-hop retrieval, self-evolution via node update operations, and on-the-fly working-memory refreshing during inference. Extensive experiments show that HyMEM consistently improves open-source GUI agents, enabling 7B/8B backbones to match or surpass strong closed-source models; notably, it boosts Qwen2.5-VL-7B by +22.5% and outperforms Gemini2.5-Pro-Vision and GPT-4o.

Wenyi Wu, Sibo Zhu, Kun Zhou et al.

Language model (LM)-based agents have demonstrated promising capabilities in automating complex tasks from natural language instructions, yet they continue to struggle with long-horizon planning and reasoning. To address this, we propose an enhanced multi-agent framework that decomposes automation into three roles: a planner for high-level decision-making, an actor for task execution, and a memory manager for contextual reasoning. While this modular decomposition aligns with established design patterns, our core contribution lies in a systematic compute-allocation analysis, revealing that planning is the dominant factor influencing task performance. Execution and memory management require significantly less compute and model capacity to achieve competitive results. Building on these insights, we introduce a planner-centric reinforcement learning approach, which exclusively optimizes the planner using trajectory-level rewards from a VLM-as-judge, while freezing the other components. Extensive experiments on benchmarks spanning web navigation, OS control, and tool use demonstrate that concentrating model capacity and learning on high-level planning yields robust and compute-efficient improvements in long-horizon agent automation. Our code is publicly released.

Wenyi Wu, Karim Bouyarmane, Ismail Tutar

We present Catalog Phrase Grounding (CPG), a model that can associate product textual data (title, brands) into corresponding regions of product images (isolated product region, brand logo region) for e-commerce vision-language applications. We use a state-of-the-art modulated multimodal transformer encoder-decoder architecture unifying object detection and phrase-grounding. We train the model in self-supervised fashion with 2.3 million image-text pairs synthesized from an e-commerce site. The self-supervision data is annotated with high-confidence pseudo-labels generated with a combination of teacher models: a pre-trained general domain phrase grounding model (e.g. MDETR) and a specialized logo detection model. This allows CPG, as a student model, to benefit from transfer knowledge from these base models combining general-domain knowledge and specialized knowledge. Beyond immediate catalog phrase grounding tasks, we can benefit from CPG representations by incorporating them as ML features into downstream catalog applications that require deep semantic understanding of products. Our experiments on product-brand matching, a challenging e-commerce application, show that incorporating CPG representations into the existing production ensemble system leads to on average 5% recall improvement across all countries globally (with the largest lift of 11% in a single country) at fixed 95% precision, outperforming other alternatives including a logo detection teacher model and ResNet50.

Wenyi Wu, Qi Li, Wenliang Zhong et al.

Vision-language models have been widely explored across a wide range of tasks and achieve satisfactory performance. However, it's under-explored how to consolidate entity understanding through a varying number of images and to align it with the pre-trained language models for generative tasks. In this paper, we propose MIVC, a general multiple instance visual component to bridge the gap between various image inputs with off-the-shelf vision-language models by aggregating visual representations in a permutation-invariant fashion through a neural network. We show that MIVC could be plugged into the visual-language models to improve the model performance consistently on visual question answering, classification and captioning tasks on a public available e-commerce dataset with multiple images per product. Furthermore, we show that the component provides insight into the contribution of each image to the downstream tasks.

Wenliang Zhong, Wenyi Wu, Qi Li et al.

Multimodal Large Language Models (MLLMs) have achieved SOTA performance in various visual language tasks by fusing the visual representations with LLMs leveraging some visual adapters. In this paper, we first establish that adapters using query-based Transformers such as Q-former is a simplified Multi-instance Learning method without considering instance heterogeneity/correlation. We then propose a general component termed Multi-instance Visual Prompt Generator (MIVPG) to incorporate enriched visual representations into LLMs by taking advantage of instance correlation between images or patches for the same sample. Quantatitive evaluation on three public vision-language (VL) datasets from different scenarios shows that the proposed MIVPG improves Q-former in main VL tasks.