Search papers, labs, and topics across Lattice.
The paper introduces Ouroboros-Spatial, a self-evolving training framework that enhances spatial reasoning in multimodal large language models (MLLMs) by allowing the model to act as both proposer and solver. This approach dynamically adjusts the training data based on the model's current capabilities, effectively filtering out trivial and overly difficult samples while generating more relevant spatial question-answer pairs. The results show significant improvements in performance on six spatial reasoning benchmarks, achieving up to 9.9 and 6.8 point gains for the 4B and 8B models, respectively, while using an order of magnitude fewer training examples than traditional datasets.
A self-evolving training framework that tailors spatial reasoning tasks to a model's current capabilities can dramatically enhance performance while minimizing data inefficiency.
Spatial reasoning remains a persistent challenge for multimodal large language models (MLLMs). Existing approaches largely rely on large-scale, statically curated datasets, where all training samples are treated uniformly regardless of the model's evolving capabilities. This static paradigm is inherently data-inefficient: training capacity is often spent on samples that are either trivial or overly difficult for the model at its current stage. To address this limitation, we propose Ouroboros-Spatial, a self-evolving training framework in which the model plays dual roles as a proposer and a solver. In each iteration, a frozen proposer generates spatial question-answer (QA) pairs from 3D scene metadata and raw video frames, together with executable code for deriving reliable ground truth. A learnable solver is then fine-tuned on the accepted samples, and its per-sample prediction confidence is used as a difficulty signal. This signal is fed back to the proposer in the next iteration, guiding it to generate questions better matched to the solver's current capabilities. Through this closed-loop design, the training distribution co-evolves with model ability, reducing redundant trivial examples while filtering out ambiguous or uninformative samples with limited learning value. Across six spatial reasoning benchmarks, Ouroboros-Spatial substantially improves Qwen3-VL-4B and Qwen3-VL-8B while using an order of magnitude fewer training examples than recent large-scale curated datasets. On VSI-Bench, it yields absolute gains of 9.9 and 6.8 points for the 4B and 8B models, respectively, enabling both to outperform a wide range of strong open-source and proprietary baselines.
