Self-evolving has emerged as a key paradigm for improving foundational models such as Large Language Models (LLMs) and Vision Language Models (VLMs) with minimal human intervention. While recent approaches have demonstrated that LLM agents can self-evolve from scratch with little to no data, VLMs introduce an additional visual modality that typically requires at least some seed data, such as images, to bootstrap the self-evolution process. In this work, we present Multi-model Multi-modal Zero (MM-Zero), the first RL-based framework to achieve zero-data self-evolution for VLM reasoning. Moving beyond prior dual-role (Proposer and Solver) setups, MM-Zero introduces a multi-role self-evolving training framework comprising three specialized roles: a Proposer that generates structured visual concepts and formulates questions; a Coder that translates these concepts into executable code (e.g., Python, SVG) to render visual images; and a Solver that performs multimodal reasoning over the generated visual content. All three roles are initialized from the same base model and trained using Group Relative Policy Optimization (GRPO), with carefully designed reward mechanisms that integrate execution feedback, visual verification, and difficulty balancing. Our experiments show that MM-Zero improves VLM reasoning performance across a wide range of multimodal benchmarks. MM-Zero establishes a scalable path toward self-evolving multi-model systems for multimodal models, extending the frontier of self-improvement beyond the conventional two-model paradigm.

Large language models (LLMs) often struggle with processing extensive input contexts, which can lead to redundant, inaccurate, or incoherent summaries. Recent methods have used unstructured memory to incrementally process these contexts, but they still suffer from information overload due to the volume of unstructured data handled. In our study, we introduce structured knowledge representations (GU_json), which significantly improve summarization performance by 40% and 14% across two public datasets. Most notably, we propose the Chain-of-Key strategy (CoK_json) that dynamically updates or augments these representations with new information, rather than recreating the structured memory for each new source. This method further enhances performance by 7% and 4% on the datasets.

Building automatic extraction models for visually rich documents like invoices, receipts, bills, tax forms, etc. has received significant attention lately. A key bottleneck in developing extraction models for new document types is the cost of acquiring the several thousand high-quality labeled documents that are needed to train a model with acceptable accuracy. In this paper, we propose selective labeling as a solution to this problem. The key insight is to simplify the labeling task to provide “yes/no” labels for candidate extractions predicted by a model trained on partially labeled documents. We combine this with a custom active learning strategy to find the predictions that the model is most uncertain about. We show through experiments on document types drawn from 3 different domains that selective labeling can reduce the cost of acquiring labeled data by 10× with a negligible loss in accuracy.

Abstract: Objective This paper aims to improve the performance of machine translation of ancient Chinese classics, which can better promote the development of ancient books research and the spread of Chinese culture. Methods Based on the multilingual translation machine pre-training model of mRASP, the model was trained by fine-tuning the specific language pairs, namely a2m, and a2e, according to the two downstream tasks of classical Chinese translation into modern Chinese and classical Chinese translation into English, using the parallel corpus of ancient white and white and ancient English parallel corpus of Pre-Qin+ZiZhiTongJian, and the translation performance of the fine-tuning model was evaluated by BIEU evaluation index. Results The BIEU4 results of the three downstream tasks of 24_histories_a2m、Pre-Qin+ZiZhiTongJian_a2m、 Pre-Qin+ZiZhiTongJian_a2e were 17.38, 13.69 and 12.90 respectively.