Diversity Collapse
No mentions found
This entity hasn't been tracked yet, or Iris is still building its knowledge base.
Related Articles from SNS
Generating Reports or Repeating Templates? Measuring and Mitigating Template Collapse in 3D CT Report Generation
Announce Type: new Abstract: Modern 3D medical vision-language models (VLMs) can generate fluent radiology-style text while exhibit critically low pathology detection and output diversity, collapsing to generic templates that under-report rare yet critical findings. We identify this failure mode as Template Collapse. This failure stems from the unique constraints of 3D medical imaging, e.g., limited data, severe label imbalance, and weak signals from volumetric encoders.
Parametric Social Identity Injection and Diversification in Public Opinion Simulation
Announce Type: replace Abstract: Large language models (LLMs) have recently been adopted as synthetic agents for public opinion simulation, offering a promising alternative to costly and slow human surveys. Despite their scalability, current LLM-based simulation methods fail to capture social diversity, producing flattened inter-group differences and overly homogeneous responses across demographic groups. We identify this limitation as a Diversity Collapse phenomenon in LLM hidden...
Climber-Pilot: A Non-Myopic Generative Recommendation Model Towards Better Instruction-Following
arXiv:2602.13581v2 Announce Type: replace Abstract: Generative retrieval has emerged as a promising paradigm in recommender systems, offering superior sequence modeling capabilities over traditional dual-tower architectures. However, in large-scale industrial scenarios, such models often suffer from inherent myopia: due to single-step inference and strict latency constraints, they tend to collapse diverse user intents into locally optimal predictions, failing to capture long-horizon and...
Initialization is Half the Battle: Generating Diverse Images from a Guidance Potential Posterior
arXiv:2606.02453v1 Announce Type: new Abstract: Despite the remarkable fidelity of generative models, they frequently suffer from mode collapse. Existing strategies for enhancing diversity predominantly focus on intervening during the generation trajectory. We identify a critical oversight that the standard Gaussian initialization often causes trajectories to collapse into dominant modes because it is agnostic to the guidance potential landscape.
When Attention Collapses: Stage-Aware Visual Token Pruning from Structure to Semantics
arXiv:2606.03569v1 Announce Type: new Abstract: Vision-Language Models (VLMs) have demonstrated remarkable capabilities but suffer from significant computational overhead during inference. While visual token pruning offers a promising solution, existing methods predominantly rely on initial attention scores. This single-metric paradigm presents a critical flaw: high attention scores inherently collapse onto semantically similar regions, thereby severely reducing feature diversity and...
Consistent Yet Wrong: Evidence Insensitivity in Spatial Vision-Language Models
arXiv:2606.02742v1 Announce Type: new Abstract: Spatial reasoning is fundamental to robotics, autonomy, and embodied AI, yet modern vision-language models (VLMs) remain unreliable on metric distance queries. A common assumption is that consistent predictions across viewpoints reflect geometric grounding.
GalaxyVS: Exploring 100-Billion Compounds in Seconds
We present GalaxyVS, a hardware-software co-designed virtual screening framework built to explore the 100-billion commercially accessible chemical space in seconds, deployed at the National Supercomputing Center in Tianjin. Built upon the dense vector retrieval paradigm of DrugCLIP, GalaxyVS bypasses the structural dependencies and computational overhead of classical docking to enable rapid screening against experimentally determined as well as geometrically feasible pockets on...
Exploiting Verification-Generation Gap: Test-Time Reinforcement Learning with Confidence-Conditioned Verification
Announce Type: new Abstract: Test-time reinforcement learning has emerged as a promising paradigm for enhancing the complex reasoning abilities of large language models in a completely label-free manner. Despite existing studies focusing on Pass@1 performance, optimizing Pass@k remains under-explored yet critical in label-free settings, which measures generation coverage for sustained exploration. Optimizing Pass@k in label-free setting is highly non-trivial, as directly applying the Pass@k...
S-SPPO: Semantic-Calibrated Self-Play Preference Optimization
arXiv:2606.01561v1 Announce Type: new Abstract: Aligning Large Language Models (LLMs) with human preferences is often formulated via Direct Preference Optimization (DPO). However, the standard Bradley-Terry instantiation of DPO is limited in modeling common departures from transitivity in human preferences. To address this, recent work has introduced Self-Play Preference Optimization (SPPO), which iteratively refines the policy by training on self-generated win-lose pairs.
"I've Seen How This Goes": Characterizing Diversity via Progressive Conditional Surprise
arXiv:2606.01811v1 Announce Type: new Abstract: Measuring the diversity of creative outputs is central to evaluating post-training mode collapse, comparing decoding strategies, and quantifying creative behavior in both AI and human writing. We propose a new approach to measuring diversity using in-context learning, of which the ``Decan'' metric, $D_{Ca_n} = C \times a_n$, is the working instance we evaluate: a per-byte score read off the per-token log-probabilities of a base model $\theta$...