inference
Inference refers to the process of using a trained AI model to make predictions or generate outputs from new input data, as opposed to training where the model learns from existing data.
- A Difference-of-Convex Functions Approach to Energy-Based Iterative Reasoning
- CAML: Collaborative Auxiliary Modality Learning for Multi-Agent Systems
- DecompNet: Enhancing Time Series Forecasting Models with Implicit Decomposition
- EMLoC: Emulator-based Memory-efficient Fine-tuning with LoRA Correction
- Efficient Prompt Compression with Evaluator Heads for Long-Context Transformer Inference
- Flow Matching Neural Processes
- Frame Context Packing and Drift Prevention in Next-Frame-Prediction Video Diffusion Models
- Learning to Focus: Causal Attention Distillation via Gradient‐Guided Token Pruning
- Lost in Latent Space: An Empirical Study of Latent Diffusion Models for Physics Emulation
- MEMOIR: Lifelong Model Editing with Minimal Overwrite and Informed Retention for LLMs
- NSNQuant: A Double Normalization Approach for Calibration-Free Low-Bit Vector Quantization of KV Cache
- Noise-Robustness Through Noise: A Framework combining Asymmetric LoRA with Poisoning MoE
- Optimal Nuisance Function Tuning for Estimating a Doubly Robust Functional under Proportional Asymptotics
- Quantifying Statistical Significance of Deep Nearest Neighbor Anomaly Detection via Selective Inference
- SING: SDE Inference via Natural Gradients
- Stochastic Forward-Forward Learning through Representational Dimensionality Compression
- Synthesize Privacy-Preserving High-Resolution Images via Private Textual Intermediaries
- Two Heads are Better than One: Simulating Large Transformers with Small Ones
- Understanding and Improving Adversarial Robustness of Neural Probabilistic Circuits
- Unified Scaling Laws for Compressed Representations
- VLMs can Aggregate Scattered Training Patches