Andi Zhang
YOU?
Author Swipe
View article: CHG combined with venetoclax and azacytidine as induction chemotherapy in acute myeloid leukemia
CHG combined with venetoclax and azacytidine as induction chemotherapy in acute myeloid leukemia Open
Background: The survival rate of patients diagnosed with acute myeloid leukemia (AML) has shown improvement in recent decades. However, the induction chemotherapy regimen for AML still relies on the standard 7+3 regimen. Standard 7+3 regim…
View article: Revoking Amnesia: RL-based Trajectory Optimization to Resurrect Erased Concepts in Diffusion Models
Revoking Amnesia: RL-based Trajectory Optimization to Resurrect Erased Concepts in Diffusion Models Open
Concept erasure techniques have been widely deployed in T2I diffusion models to prevent inappropriate content generation for safety and copyright considerations. However, as models evolve to next-generation architectures like Flux, establi…
View article: Rethinking Inter-LoRA Orthogonality in Adapter Merging: Insights from Orthogonal Monte Carlo Dropout
Rethinking Inter-LoRA Orthogonality in Adapter Merging: Insights from Orthogonal Monte Carlo Dropout Open
We propose Orthogonal Monte Carlo Dropout, a mechanism that enforces strict orthogonality when combining sparse semantic vectors without extra time complexity. Low-Rank Adaptation (LoRA), a popular fine-tuning method for large models, typi…
View article: Concept-based Adversarial Attack: a Probabilistic Perspective
Concept-based Adversarial Attack: a Probabilistic Perspective Open
We propose a concept-based adversarial attack framework that extends beyond single-image perturbations by adopting a probabilistic perspective. Rather than modifying a single image, our method operates on an entire concept -- represented b…
View article: Compositional Attribute Imbalance in Vision Datasets
Compositional Attribute Imbalance in Vision Datasets Open
Visual attribute imbalance is a common yet underexplored issue in image classification, significantly impacting model performance and generalization. In this work, we first define the first-level and second-level attributes of images and t…
View article: D56. Large Language Models for Perceptual Speech Clinical Data Extraction in 50 Cleft Palate Speech Language Pathology Notes
D56. Large Language Models for Perceptual Speech Clinical Data Extraction in 50 Cleft Palate Speech Language Pathology Notes Open
View article: 48. Large Language Models for Perceptual Speech Clinical Data Extraction in Cleft Palate Speech Language Pathology Notes
48. Large Language Models for Perceptual Speech Clinical Data Extraction in Cleft Palate Speech Language Pathology Notes Open
View article: Assessment of Union Rates in Ballistic Hand Fractures
Assessment of Union Rates in Ballistic Hand Fractures Open
Background: Ballistic hand fractures (BHFs) are complex, multistructure injuries, unique from other hand fractures. We sought to determine BHF union rates and assess factors contributing to nonunion. Methods: Metacarpal or phalangeal BHFs …
View article: Recycling Revolution or Environmental Setback? Unveiling the Rebound Effects of Metal Recycling Technological Innovation
Recycling Revolution or Environmental Setback? Unveiling the Rebound Effects of Metal Recycling Technological Innovation Open
View article: Multifunctional Organic Aniline Molecule with Interface and Bulk Defect Modification for Efficient Inverted Perovskite Solar Cells
Multifunctional Organic Aniline Molecule with Interface and Bulk Defect Modification for Efficient Inverted Perovskite Solar Cells Open
View article: Noise Diffusion for Enhancing Semantic Faithfulness in Text-to-Image Synthesis
Noise Diffusion for Enhancing Semantic Faithfulness in Text-to-Image Synthesis Open
Diffusion models have achieved impressive success in generating photorealistic images, but challenges remain in ensuring precise semantic alignment with input prompts. Optimizing the initial noisy latent offers a more efficient alternative…
View article: LoKO: Low-Rank Kalman Optimizer for Online Fine-Tuning of Large Models
LoKO: Low-Rank Kalman Optimizer for Online Fine-Tuning of Large Models Open
Training large models with millions or even billions of parameters from scratch incurs substantial computational costs. Parameter Efficient Fine-Tuning (PEFT) methods, particularly Low-Rank Adaptation (LoRA), address this challenge by adap…
View article: DIG-FACE: De-biased Learning for Generalized Facial Expression Category Discovery
DIG-FACE: De-biased Learning for Generalized Facial Expression Category Discovery Open
We introduce a novel task, Generalized Facial Expression Category Discovery (G-FACE), that discovers new, unseen facial expressions while recognizing known categories effectively. Even though there are generalized category discovery method…
View article: Prediction-based multi-objective optimization method for 3D printing resource consumption
Prediction-based multi-objective optimization method for 3D printing resource consumption Open
A prediction-based multi-objective optimization (PBMO) method is proposed in this paper to forecast and reduce 3D printing (3DP) resources on demand, including time, energy, and material. In the authors’ previous research work, a hybrid co…
View article: Improving Probabilistic Diffusion Models With Optimal Diagonal Covariance Matching
Improving Probabilistic Diffusion Models With Optimal Diagonal Covariance Matching Open
The probabilistic diffusion model has become highly effective across various domains. Typically, sampling from a diffusion model involves using a denoising distribution characterized by a Gaussian with a learned mean and either fixed or le…
View article: Public Perception of Scrub Color and Style in Plastic Surgery
Public Perception of Scrub Color and Style in Plastic Surgery Open
Background: Scrubs have become widespread office attire for plastic surgeons. The purpose of this study is to evaluate the public perception of scrub color and style for plastic surgeons. Methods: A crowdsourced survey was performed via MT…
View article: Recurrent Ileocolic Intussusception With the Appendix as the Pathologic Lead Point in Children: A Report of Two Cases and Review of Literature
Recurrent Ileocolic Intussusception With the Appendix as the Pathologic Lead Point in Children: A Report of Two Cases and Review of Literature Open
Ileocolic intussusception is a consideration in young pediatric patients with acute abdominal pain. Meckel's diverticulum is the most common pathologic lead point for intussusception in children and the appendix acting as the lead point is…
View article: A novel homozygous RSPH4A variant in a family with primary ciliary dyskinesia and literature review
A novel homozygous RSPH4A variant in a family with primary ciliary dyskinesia and literature review Open
Introduction: Primary ciliary dyskinesia (PCD) is a rare heterogeneous disease caused by abnormalities in motile cilia. In this case report, we first analyzed the clinical and genetic data of a proband who was suspected of having PCD on th…
View article: Your Finetuned Large Language Model is Already a Powerful Out-of-distribution Detector
Your Finetuned Large Language Model is Already a Powerful Out-of-distribution Detector Open
We revisit the likelihood ratio between a pretrained large language model (LLM) and its finetuned variant as a criterion for out-of-distribution (OOD) detection. The intuition behind such a criterion is that, the pretrained LLM has the pri…
View article: 57. Challenging The Paradigm: Rates Of Mandibular Infection Following Fracture
57. Challenging The Paradigm: Rates Of Mandibular Infection Following Fracture Open
Purpose: Mandible fractures are common in trauma and account for the majority of adult facial fracture reductions in the United States. Infectious complications including wound or hardware infections, osteomyelitis, and abscess formation a…
View article: Prediction-based multi-objective optimisation method for 3D Printing Resource Consumption
Prediction-based multi-objective optimisation method for 3D Printing Resource Consumption Open
A prediction-based multi-objective optimisation (PBMO) method is proposed for 3D printing (3DP) technology to predict and reduce resource requirements on-demand, including time, energy and material. In the authors’ previous research work, …
View article: Impact of cholesterol homeostasis within cochlear cells on auditory development and hearing loss
Impact of cholesterol homeostasis within cochlear cells on auditory development and hearing loss Open
Cholesterol is the most abundant sterol molecule in mammalian cells, which not only constitutes the cell membrane but also plays essential roles in the synthesis of important hormones, synapse formation, and cell signal transduction. The e…
View article: "Lost in Translation: The Readability Discrepancy of Online Patient Educational Materials for PCL Surgery".
"Lost in Translation: The Readability Discrepancy of Online Patient Educational Materials for PCL Surgery". Open
There is a discrepancy in readability between the recommendation of AMA/NIH and online educational materials regarding PCL surgeries, regardless of where they appear on Google and across different forums. The understandability and actionab…
View article: Speech Analysis of Patients with Cleft Palate Using Artificial Intelligence Techniques: A Systematic Review
Speech Analysis of Patients with Cleft Palate Using Artificial Intelligence Techniques: A Systematic Review Open
Introduction: This systematic review examines the use of machine learning (ML) algorithms to detect hypernasal speech in patients with cleft palate (CP), which can persist after primary repair surgery, and require revision. Due to a shorta…
View article: Constructing Semantics-Aware Adversarial Examples with a Probabilistic Perspective
Constructing Semantics-Aware Adversarial Examples with a Probabilistic Perspective Open
We propose a probabilistic perspective on adversarial examples, allowing us to embed subjective understanding of semantics as a distribution into the process of generating adversarial examples, in a principled manner. Despite significant p…
View article: SR-OOD: Out-of-Distribution Detection via Sample Repairing
SR-OOD: Out-of-Distribution Detection via Sample Repairing Open
Out-of-distribution (OOD) detection is a crucial task for ensuring the reliability and robustness of machine learning models. Recent works have shown that generative models often assign high confidence scores to OOD samples, indicating tha…
View article: Spatio-temporal pattern of c-Jun N-terminal kinase isoforms in the cochleae of C57BL/6J mice with presbycusis
Spatio-temporal pattern of c-Jun N-terminal kinase isoforms in the cochleae of C57BL/6J mice with presbycusis Open
The c-Jun N-terminal kinase (JNK) pathway is a vital component of the mitogen-activated protein kinase cascade, which regulates cell death and survival. The present study aimed to explore the Spatio-temporal changes in all JNK isoforms in …
View article: An Extreme Case of Lethal Abdominal Compartment Syndrome in Pediatric Patient With Synovial Sarcoma
An Extreme Case of Lethal Abdominal Compartment Syndrome in Pediatric Patient With Synovial Sarcoma Open
View article: Falsehoods that ML researchers believe about OOD detection
Falsehoods that ML researchers believe about OOD detection Open
An intuitive way to detect out-of-distribution (OOD) data is via the density function of a fitted probabilistic generative model: points with low density may be classed as OOD. But this approach has been found to fail, in deep learning set…
View article: Impacts of impaired mitochondrial dynamics in hearing loss: Potential therapeutic targets
Impacts of impaired mitochondrial dynamics in hearing loss: Potential therapeutic targets Open
Mitochondria are the powerhouse of the cells. Under physiological conditions, mitochondrial fission and fusion maintain a dynamic equilibrium in the cytoplasm, which is referred to as mitochondrial dynamics. As an important approach to reg…