REVIEW 8 cited by
Set You Straight: Auto-Steering Denoising Trajectories to Sidestep Unwanted Concepts
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Ensuring the ethical deployment of text-to-image models requires effective techniques to prevent the generation of harmful or inappropriate content. While concept erasure methods offer a promising solution, existing finetuning-based approaches suffer from notable limitations. Anchor-free methods risk disrupting sampling trajectories, leading to visual artifacts, while anchor-based methods rely on the heuristic selection of anchor concepts. To overcome these shortcomings, we introduce a finetuning framework, dubbed ANT, which Automatically guides deNoising Trajectories to avoid unwanted concepts. ANT is built on a key insight: reversing the condition direction of classifier-free guidance during mid-to-late denoising stages enables precise content modification without sacrificing early-stage structural integrity. This inspires a trajectory-aware objective that preserves the integrity of the early-stage score function field, which steers samples toward the natural image manifold, without relying on heuristic anchor concept selection. For single-concept erasure, we propose an augmentation-enhanced weight saliency map to precisely identify the critical parameters that most significantly contribute to the unwanted concept, enabling more thorough and efficient erasure. For multi-concept erasure, our objective function offers a versatile plug-and-play solution that significantly boosts performance. Extensive experiments demonstrate that ANT achieves state-of-the-art results in both single and multi-concept erasure, delivering high-quality, safe outputs without compromising the generative fidelity. Code is available at https://github.com/lileyang1210/ANT
Forward citations
Cited by 8 Pith papers
-
Decoupled Spatio-Temporal Consistency Learning for Self-Supervised Tracking
SSTrack trains a Vision Transformer tracker without frame-wise box labels by combining forward global search, backward local association, and instance contrastive learning, and reports state-of-the-art self-supervised...
-
Show and Polish: Reference-Guided Identity Preservation in Face Video Restoration
IP-FVR restores degraded face videos with consistent identity by conditioning a video diffusion model on a reference photo of the same person.
-
CRISP-SAM2: SAM2 with Cross-Modal Interaction and Semantic Prompting for Multi-Organ Segmentation
A text-guided SAM2 variant with cross-modal attention, semantic prompt generation, and a similarity-sorted memory bank achieves top Dice and surface scores on seven public multi-organ CT datasets.
-
Enhancing Zero-Shot Brain Tumor Subtype Classification via Fine-Grained Patch-Text Alignment
FG-PAN improves zero-shot brain tumor subtype classification by aligning refined visual patch features with LLM-generated fine-grained text prototypes.
-
FADE: Adversarial Concept Erasure in Flow Models
FADE combines adversarial training with trajectory preservation to erase concepts from diffusion models, reporting state-of-the-art erasure on Stable Diffusion benchmarks, but the evidence is incomplete and the theore...
-
Erasing Concepts, Steering Generations: A Comprehensive Survey of Concept Suppression
This survey classifies concept erasure methods for text-to-image diffusion models along intervention level, optimization strategy, and semantic scope, and reviews the datasets, metrics, and benchmarks used to evaluate them.
-
Butter: Frequency Consistency and Hierarchical Fusion for Autonomous Driving Object Detection
A YOLO-family detector named Butter claims state-of-the-art efficiency on KITTI, BDD100K, and Cityscapes, but the paper's loss equations, parameter counts, and baseline comparisons contain contradictions that undermin...
-
TRACE: Trajectory-Constrained Concept Erasure in Diffusion Models
TRACE combines a closed-form cross-attention nullification with a late-timestep fine-tuning loss to erase concepts from diffusion models, claiming better erasure and fidelity than published baselines.
Discussion (0). Sign in to comment.