Newest papers

Spatially Gated Demographic Steering for Debiasing Multi-Person Text-to-Image Generation

Venkatesh Thirugnana Sambandham, Torsten Schön · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.29651 · cited by 0 works (OpenAlex) · open-access PDF

Indonesian

Model teks ke gambar mereproduksi stereotip masyarakat, dan mitigasi yang ada menerapkan satu koreksi demografis ke seluruh adegan. Dalam generasi adegan multi-pribadi ini menciptakan kompromi: pengendalian global meningkatkan keseimbangan agregat hanya dengan homogenisasi orang-orang dalam gambar. Metode berbasis pengendalian seperti Embedding Arithmetic mendorong keragaman intra-image di bawah garis awal yang tidak terkontrol. Pendekatan teks dan proyeksi latent mengalihkan gesekan agregat dengan hanya keuntungan marginal dalam keragaman intra-image. Kami menyajikan Spatially Gated Steering (SGS), yang mengendalikan demografi setiap orang secara independen dengan mengekstrak masker per orang dari peta perhatian internal Transformer (DiT) pada waktu denoising kritis dan menerapkan vektor demografis yang berbeda dalam setiap masker untuk langkah

Machine translation (Meta M2M100); the English original follows.

Abstract

Text-to-Image models reproduce societal stereotypes, and existing mitigation applies one demographic correction to the whole scene. In multi-person scene generation this creates a trade-off: global steering improves aggregate balance only by homogenizing the people within an image. Steering-based methods like Embedding Arithmetic drive the intra-image diversity below the unsteered baseline. Text- and latent-projection approaches shift the aggregate skew with only marginal gains in intra-image diversity. We present Spatially Gated Steering (SGS), which steers the demographics of each person independently by extracting per-person masks from the Diffusion Transformer's (DiT) internal attention maps at a critical denoising timestep and applying a distinct demographic vector within each mask for the remaining steps. SGS needs no fine-tuning, inversion or layout specifications, and uses a lightweight segmentor over frozen DiT attention features to localize subjects at early timesteps. On skewed occupational prompts, SGS reaches high intra-image racial and gender diversity while keeping aggregate balance competitive and image quality near baseline, a result we validate on two architecturally distinct diffusion transformers, FLUX.1-Dev and SD3.5-Large. Codebase and model weights are available at https://github.com/cvims/spatially-gated-steering

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Debiasing, Skew, Computer science, Glyph (data visualization), Aggregate (composite), Offset (computer science)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange