Диффузионные методы генерации, дополнения или исправления программ. DiffuCoder изучает итеративное восстановление кода и варианты обучения; перед интерпретацией оценки определите контракт задачи.
Как читать и сравнивать
Это дополнение, преобразование, исправление или заполнение пропуска и какой контекст должен сохраниться?
Отдельно ли измерены синтаксическая корректность, исполнение или тесты и функциональная правильность?
Исследование генерации кода не гарантирует правильные или безопасные программы либо общее преимущество над другими способами генерации.
Наличие кода не гарантирует воспроизводимость. Сверяйте набор данных, метрику, протокол и условия в оригинале.
Оригинальная аннотация: Masked diffusion language models predict tokens from a partially observed response canvas, enabling bidirectional conditioning and parallel token refinement. Yet standard masked-diffusion decoders use a rigid inference interface: the number of masked positions allocated to the answer is fixed before generation begins. Choosing this length is difficult. A short canvas can truncate reasoning or code, while a long canvas wastes computation and can perturb denoising. We introduce CARVE (Counterfactual-Aware Reveal with Verified Expansion), a training-free variable-length algorithm for masked diffu
Оригинальная аннотация: MDLMs generate text by denoising a preallocated masked response canvas, making response-length modeling central to instruction tuning. Existing MDLMs often inherit the autoregressive convention of using repeated \texttt{[EOS]} tokens for padding during instruction tuning, giving \texttt{[EOS]} a dual role as both a semantic terminator and a padding token. We show that this dual role is a root cause of \texttt{[EOS]} overflow under large-block decoding. To decouple these roles, we propose VoidPadding, which introduces \texttt{[VOID]} for padding and reserves \texttt{[EOS]} for termination. Duri
Оригинальная аннотация: Block-wise semi-autoregressive decoding is the standard inference paradigm for diffusion large language models (DLMs), but it imposes a strict dependency between blocks: the next block cannot begin until the current block is fully decoded or its denoising budget is exhausted. We observe that once a block exposes a reliable delimiter boundary or stable semantic prefix, continuation generation need not wait for every residual token to be resolved. We propose AsyncLane, a training-free decoding scheduler that decouples refinement from advancement. AsyncLane forks a generate lane at observed delim
Полезно для: Не указано · Ограничение: Не указано
2026-05-16T06:15:47+00:00 · Маскированная / дискретная диффузия, Генерация кода · Источник
Оригинальная аннотация: Discrete diffusion models are a powerful, emerging paradigm for code generation. They construct programs through iterative refinement of partially corrupted token sequences and enable parallel token refinement. Importantly, this paradigm exposes a global program state at each denoising step, which provides a natural intervention point for enforcing program-level functionality and security constraints, guiding the generation before the final code is committed. Building on this observation, the paper introduces Constrained Diffusion for Code (CDC), a training-free neurosymbolic inference framewo