12 papers
Test-Time Coverage: Test-Conditioned Data Curation for Deployment-Aware Learning
Nadine Chang, Maying Shen, Shizhe Diao +6
Deployed AI systems are often trained from broad candidate data pools, necessitating data curation towards the deployment test distribution. However, standard data curation methods…
From Modalities to Propositions: A Language-Centric Framework for Multimodal Intelligence
Nadine Chang, Maying Shen, Shizhe Diao +6
We propose a language representation for multimodal data in which any observation, whether image, video, or text, is expressed as a bag of atomic propositions, simple statements ab…
Position: Stop Reactively Patching Your Model Every Time and Start Proactive Test-Driven AI Development
Nadine Chang, Maying Shen, Jialiang Wang +2
Many modern AI systems are designed to operate under diverse, open-ended, use-cases. To help generalize deployed systems, many deployed-system maintenance pipelines use a reactive…
Scaling-Aware Data Selection for End-to-End Autonomous Driving Systems
Tolga Dimlioglu, Nadine Chang, Maying Shen +2
Large-scale deep learning models for physical AI applications depend on diverse training data collection efforts. These models and correspondingly, the training data, must address…
DriveCritic: Towards Context-Aware, Human-Aligned Evaluation for Autonomous Driving with Vision-Language Models
Jingyu Song, Zhenxin Li, Shiyi Lan +6
Benchmarking autonomous driving planners to align with human judgment remains a critical challenge, as state-of-the-art metrics like the Extended Predictive Driver Model Score (EPD…
Mitigating Multimodal Hallucinations via Gradient-based Self-Reflection
Shan Wang, Maying Shen, Nadine Chang +3
Multimodal large language models achieve strong performance across diverse tasks but remain prone to hallucinations, where outputs are not grounded in visual inputs. This issue can…
PARC: A Quantitative Framework Uncovering the Symmetries within Vision Language Models
Jenny Schmalfuss, Nadine Chang, Vibashan VS +3
Vision language models (VLMs) respond to user-crafted text prompts and visual inputs, and are applied to numerous real-world problems. VLMs integrate visual modalities with large l…
MDP: Multidimensional Vision Model Pruning with Latency Constraint
Xinglong Sun, Barath Lakshmanan, Maying Shen +3
Current structural pruning methods face two significant limitations: (i) they often limit pruning to finer-grained levels like channels, making aggressive parameter reduction chall…
Advancing Weight and Channel Sparsification with Enhanced Saliency
Xinglong Sun, Maying Shen, Hongxu Yin +3
Pruning aims to accelerate and compress models by removing redundant parameters, identified by specifically designed importance scores which are usually imperfect. This removal is…
Global Average Feature Augmentation for Robust Semantic Segmentation with Transformers
Alberto Gonzalo Rodriguez Salgado, Maying Shen, Philipp Harzig +2
Robustness to out-of-distribution data is crucial for deploying modern neural networks. Recently, Vision Transformers, such as SegFormer for semantic segmentation, have shown impre…
SSE: Multimodal Semantic Data Selection and Enrichment for Industrial-scale Data Assimilation
Maying Shen, Nadine Chang, Sifei Liu +1
In recent years, the data collected for artificial intelligence has grown to an unmanageable amount. Particularly within industrial applications, such as autonomous vehicles, model…
Step Out and Seek Around: On Warm-Start Training with Incremental Data
Maying Shen, Hongxu Yin, Pavlo Molchanov +2
Data often arrives in sequence over time in real-world deep learning applications such as autonomous driving. When new training data is available, training the model from scratch u…