MM-Eval unifies evaluation of multimodal summaries by integrating factual text quality, cross-modal relevance via MLLM judge, and visual diversity via truncated CLIP entropy, then calibrates their combination on human preferences.
hub
Weinberger , editor =
14 Pith papers cite this work, alongside 130 external citations. Polarity classification is still indexing.
14
Pith papers citing it
130
external citations · OpenAlex
hub tools
citation-role summary
method 1
citation-polarity summary
roles
method 1polarities
use method 1representative citing papers
MIRL uses mutual information to guide trajectory selection and provide separate rewards for visual perception in RLVR for VLMs, achieving 70.22% average accuracy with 25% fewer full trajectories.
LLM.int8() performs 8-bit inference for transformers up to 175B parameters with no accuracy loss by combining vector-wise quantization for most features with 16-bit mixed-precision handling of systematic outlier dimensions.
Quantized reasoning models produce longer chains of thought, inflating token usage and negating per-token speedups from low-bit quantization across multiple benchmarks.
Introduces a representation-geometry-based taxonomy for continual learning in speech and audio, identifies mismatches with current CL assumptions in foundation models, and lists open challenges.
Mean-field theory of dropout at the edge of chaos derives scaling laws showing front-loaded schedules outperform constant dropout by shifting the perfect-alignment fixed point.
SPeCTrA-Sum uses hierarchical cross-modal fusion via DVP and DPP-distilled image selection via VRP to generate more accurate and visually grounded multimodal summaries.
RAT estimates Tikhonov-regularized natural policy gradients by rewriting them with the Woodbury identity, approximating the transformed advantage via randomized block Kaczmarz, and applying it as a vanilla policy gradient surrogate.
ViSA-R2 recovers single executable SymPy expressions for linear steady-state fields from visualizations using a self-verifying chain-of-thought that recognizes patterns, hypothesizes solution families, derives parameters, and checks consistency.
VCON is a unified framework for smooth iterative DNN compression that uses parallel execution and an affine combination to progressively replace the original model with its compressed form during fine-tuning.
MAPE combines a channel-attention U-Net (SAPE) trained on multi-model adversarial examples scheduled by PPSA to eliminate perturbations, reporting over 95.1% average defense on CIFAR-10 and 71.5% on Mini-ImageNet against black-box transferable attacks.
KLR Hopfield networks exhibit robustness to quantization but sensitivity to pruning, interpreted as arising from dense bimodal parameterization of sparse input mappings.
Gated-SwinRMT unifies Swin windowed attention with retentive Manhattan decay via gating, reaching 80.22% top-1 accuracy on Mini-ImageNet versus 73.74% for the RMT baseline.
A DenseNet201 base model trained on a constructed plant leaf disease dataset outperforms baselines and enables faster, more robust transfer learning with less data than general models.
citing papers explorer
-
Measuring What Matters Beyond Text: Evaluating Multimodal Summaries by Quality, Alignment, and Diversity
MM-Eval unifies evaluation of multimodal summaries by integrating factual text quality, cross-modal relevance via MLLM judge, and visual diversity via truncated CLIP entropy, then calibrates their combination on human preferences.
-
MIRL: Mutual Information-Guided Reinforcement Learning for Vision-Language Models
MIRL uses mutual information to guide trajectory selection and provide separate rewards for visual perception in RLVR for VLMs, achieving 70.22% average accuracy with 25% fewer full trajectories.
-
LLM.int8(): 8-bit Matrix Multiplication for Transformers at Scale
LLM.int8() performs 8-bit inference for transformers up to 175B parameters with no accuracy loss by combining vector-wise quantization for most features with 16-bit mixed-precision handling of systematic outlier dimensions.
-
Quantization Inflates Reasoning: Token Inflation as a Hidden Cost of Low-Bit Reasoning Models
Quantized reasoning models produce longer chains of thought, inflating token usage and negating per-token speedups from low-bit quantization across multiple benchmarks.
-
Rethinking Continual Learning for Speech and Audio: A Representation-Centric Taxonomy and Open Problems
Introduces a representation-geometry-based taxonomy for continual learning in speech and audio, identifies mismatches with current CL assumptions in foundation models, and lists open challenges.
-
Dropout Universality: Scaling Laws and Optimal Scheduling at the Edge-of-Chaos
Mean-field theory of dropout at the edge of chaos derives scaling laws showing front-loaded schedules outperform constant dropout by shifting the perfect-alignment fixed point.
-
Towards Visually Grounded Multimodal Summarization via Cross-Modal Transformer and Gated Attention
SPeCTrA-Sum uses hierarchical cross-modal fusion via DVP and DPP-distilled image selection via VRP to generate more accurate and visually grounded multimodal summaries.
-
Randomized Advantage Transformation (RAT): Computing Natural Policy Gradients via Direct Backpropagation
RAT estimates Tikhonov-regularized natural policy gradients by rewriting them with the Woodbury identity, approximating the transformed advantage via randomized block Kaczmarz, and applying it as a vanilla policy gradient surrogate.
-
Hidden in Plain Sight: Visual-to-Symbolic Analytical Solution Inference from Field Visualizations
ViSA-R2 recovers single executable SymPy expressions for linear steady-state fields from visualizations using a self-verifying chain-of-thought that recognizes patterns, hypothesizes solution families, derives parameters, and checks consistency.
-
Vanishing Contributions: A Unified Framework for Smooth and Iterative Model Compression
VCON is a unified framework for smooth iterative DNN compression that uses parallel execution and an affine combination to progressively replace the original model with its compressed form during fine-tuning.
-
MAPE: Defending Against Transferable Adversarial Attacks Using Multi-Source Adversarial Perturbations Elimination
MAPE combines a channel-attention U-Net (SAPE) trained on multi-model adversarial examples scheduled by PPSA to eliminate perturbations, reporting over 95.1% average defense on CIFAR-10 and 71.5% on Mini-ImageNet against black-box transferable attacks.
-
Quantization robustness from dense representations of sparse functions in high-capacity kernel associative memory
KLR Hopfield networks exhibit robustness to quantization but sensitivity to pruning, interpreted as arising from dense bimodal parameterization of sparse input mappings.
-
Gated-SwinRMT: Unifying Swin Windowed Attention with Retentive Manhattan Decay via Input-Dependent Gating
Gated-SwinRMT unifies Swin windowed attention with retentive Manhattan decay via gating, reaching 80.22% top-1 accuracy on Mini-ImageNet versus 73.74% for the RMT baseline.
-
Developing a Strong Pre-Trained Base Model for Plant Leaf Disease Classification
A DenseNet201 base model trained on a constructed plant leaf disease dataset outperforms baselines and enables faster, more robust transfer learning with less data than general models.