Scalable and Efficient Mamba like Attention for computer vision tasks.

A situated LLM architecture separating cognitive and executive layers for stable emotional support interaction.

Retrieval High-quAlity Demonstrations for decision-making in offline RL.

A GNN approach utilizing argument mining to detect LLM authorship more robustly than linguistic features.

Synthetic face datasets used for privacy-preserving face recognition training and benchmarking.

A distillation-based pretraining framework for Multiple Instance Learning networks.

End-to-end framework for apparent personality inference from facial images.

Code for a multi-species plant identification pipeline using DINOv2 and FAISS.

Repository for multi-label animal vocalization detection in soundscapes.

A function-based edge computing framework for heterogeneous resource utilization.

A surrogate model for laminated-core material based on recurrent neural networks for finite element simulations.

Benchmark for evaluating MLLM literacy in scientific visualizations.

A training framework for efficient multimodal document question answering using GRPO.

A benchmark comprising on-ball decisions from the FIFA World Cup to measure strategic reasoning in VLMs.

A framework for improving Vision-Language-Action (VLA) models via structured representation shaping.

A self-evolving topological agent framework for multimodal scientific reasoning.

OpenTelemetry-based governance stack for AI alignment and regulatory compliance.

Source code and data for classifying cuneiform tablet metadata using 3D point clouds.

Framework for mitigating class imbalance in graph node classification using node importance assessment.

The navigation stack used to execute robotic movement goals based on semantic perception.