The Spatial Blindspot of Vision-Language Models
Paper • 2601.09954 • Published
AI & ML interests
Agents, Reasoning, Evals
Recent Activity
View all activity
ignoreandfly
authored 2
papers 3 days ago
raghavprabhakar
authored 2
papers 25 days ago
patrickamadeus
authored 5
papers about 2 months ago
Seeing Culture: A Benchmark for Visual Reasoning and Grounding
Paper • 2509.16517 • Published • 3
Vision Language Models are Confused Tourists
Paper • 2511.17004 • Published • 1
Can Large Language Models Understand, Reason About, and Generate Code-Switched Text?
Paper • 2601.07153 • Published
M4-RAG: A Massive-Scale Multilingual Multi-Cultural Multimodal RAG
Paper • 2512.05959 • Published
LinguDistill: Recovering Linguistic Ability in Vision- Language Models via Selective Cross-Modal Distillation
Paper • 2604.00829 • Published • 8
patrickamadeus
submitted a
paper to Daily Papers about 2 months ago
syedmohaiminulhoque
authored a
paper 2 months ago
Akhil-Theerthala
submitted a
paper to Daily Papers 5 months ago
Akhil-Theerthala
authored a
paper 5 months ago
Akhil-Theerthala
posted an update 5 months ago
Post
483
Is it better to show a model too many images once (Diversity), or extract as much information from a small set of images?
I have always wanted to do an ablation study on this and recently I got the chance to do exactly that. Why? In applied domains like robotics, manufacturing, or banking, we rarely have the luxury of internet-scale diverse image datasets. We are often "Data Poor" in terms of diversity but "Data Rich" in depth.
The takeaway? Density is efficient for facts but dangerous for reasoning (logical collapse) if you don't have larger scale data.
More details:
https://huggingface.co/blog/Akhil-Theerthala/diversity-density-for-vision-language-models
I have always wanted to do an ablation study on this and recently I got the chance to do exactly that. Why? In applied domains like robotics, manufacturing, or banking, we rarely have the luxury of internet-scale diverse image datasets. We are often "Data Poor" in terms of diversity but "Data Rich" in depth.
The takeaway? Density is efficient for facts but dangerous for reasoning (logical collapse) if you don't have larger scale data.
More details:
https://huggingface.co/blog/Akhil-Theerthala/diversity-density-for-vision-language-models
- 1 reply
smsnobin77
authored 5
papers 6 months ago
FedMentor: Domain-Aware Differential Privacy for Heterogeneous Federated LLMs in Mental Health
Paper • 2509.14275 • Published
FedMentalCare: Towards Privacy-Preserving Fine-Tuned LLMs to Analyze Mental Health Status Using Federated Learning Framework
Paper • 2503.05786 • Published
FilterRAG: Zero-Shot Informed Retrieval-Augmented Generation to Mitigate Hallucinations in VQA
Paper • 2502.18536 • Published
Differentially Private Data Publication with Multi-level Data Utility
Paper • 2112.07061 • Published
Differential Privacy in Privacy-Preserving Big Data and Learning: Challenge and Opportunity
Paper • 2112.01704 • Published
davidquicast
posted an update 6 months ago
Post
4289
- 3 replies