ELF improves FM forecasts by efficiently using online feedback.
problem Fixed FMs fail to adapt to current data characteristics.
method ELF-Forecaster learns current data distribution, ELF-Weighter combines forecasts.
result ELF improves FM forecasts across various datasets.
Mapper-GIN simplifies 3D point cloud classification with lightweight structure.
problem Robust 3D point cloud classification under corruption.
method Mapper algorithm for structural decomposition, GIN for graph classification.
result Mapper-GIN achieves competitive accuracy with minimal parameters.
GIBLy adds geometric priors to 3D segmentation models, improving performance with minimal overhead.
problem Lack of explicit geometric information in 3D semantic segmentation models.
method Introduces GIBLy, a lightweight geometric inductive bias layer that integrates learnable geometric priors into existing 3D segmentation pipelines.
result Consistent performance gains across multiple benchmarks, including up to +11.5% mIoU on TS40K with PTV3.
Paper proposes FMore to incentivize edge nodes in federated learning with MEC.
problem Incentivizing edge nodes in federated learning with MEC resources.
method Multi-dimensional procurement auction with K winners.
result FMore improves model accuracy and reduces training rounds for AI tasks.
TinyXRA assesses financial risks from 10-K reports using a lightweight transformer model.
problem Comprehensive risk assessment from financial reports, distinguishing between upside and downside risk.
method Lightweight transformer model with dynamic attention, incorporating skewness, kurtosis, and Sortino ratio.
result State-of-the-art predictive accuracy and transparent risk assessments.
Finet uses FBN for efficient, lightweight neural networks.
problem Building efficient neural networks with limited computational resources.
method Introduces Fine-grained Batch Normalization (FBN) and a novel light-weight network (Finet) that combines FBN with standard convolution.
result Finet achieves state-of-the-art performance on ImageNet classification with reduced computational complexity.
A new framework for lightweight BNNs learns heteroscedastic uncertainties efficiently.
problem Learning heteroscedastic uncertainties from BNNs for lightweight networks.
method Embedding heteroscedastic variances into BNN parameters and using moment propagation for inference.
result Improves predictive performance for lightweight BNNs without increasing parameter count.
DBQ quantizes lightweight networks efficiently for resource-constrained devices.
problem High computational and storage complexity of deep neural networks on resource-constrained devices.
method A differentiable non-uniform quantizer that can be mapped onto efficient ternary-based dot product engines.
result Achieves state-of-the-art results with minimal training overhead and best accuracy-complexity trade-off.
RSAC improves lightweight continuous learning efficiency.
problem Excessive training time and memory usage in continuous learning.
method Regularized subspace approximation classifier with feature reduction and regularization.
result RSAC achieves more efficient continuous learning than prior methods.
DLC enhances distillation-based continual learning with lightweight plugins.
problem Stability-plasticity dilemma in distillation-based continual learning.
method DLC deploys lightweight residual plugins into the classifier-proximal layer of a shared feature extractor.
result Significant 8% accuracy gain on large-scale benchmarks with minimal parameter increase.
iSTFTNet2 improves iSTFTNet's speed and lightness with 1D-2D CNN.
problem Efficiently synthesizing high-fidelity speech.
method Improved iSTFTNet using 1D-2D CNNs for temporal and spectrogram structures.
result iSTFTNet2 is faster and more lightweight with comparable speech quality.
Augment small datasets with synthetic backgrounds to train lightweight CNNs for human pose estimation.
problem Training CNNs from limited real-world data for human pose estimation.
method Synthetic background substitution for data augmentation.
result Improves generalization to unseen environments.
We present PROPS, a lightweight transfer learning mechanism for sequential data. PROPS learns probabilistic perturbations around the predictions of one or more arbitrarily complex, pre-trained black box models (such as recurrent neural networks). The technique pins the black-box prediction functions to "source nodes" o…
CMoS improves time series forecasting with minimal parameters.
problem Efficiently forecasting time series data with limited resources.
method CMoS directly models chunk-wise spatial correlations, using Correlation Mixing and Periodicity Injection techniques.
result CMoS outperforms state-of-the-art models with minimal parameters.
Federated learning framework improves model generalization and privacy.
problem Communication overhead and statistical heterogeneity in FL.
method Prototypes and lightweight adapters for local model refinement.
result Improves classification accuracy over baseline algorithms.
FastGAT reduces GNN computation time by 10x using graph sparsification.
problem High computational burden in attention-based GNNs.
method Spectral sparsification to generate optimal graph pruning.
result Per-epoch time is almost linear in graph nodes, reducing computational time by up to 10x.
SFBoW provides sentence embeddings with predefined dimensions.
problem Sentence embeddings problem at document-level.
method Refinement of Fuzzy Bag-of-Words, predefined dimension.
result Competitive performances in Semantic Textual Similarity benchmarks.
BERTino is a lightweight Italian DistilBERT model for NLP tasks.
problem High computational and memory demands of large language models.
method Developed a DistilBERT model tailored for Italian language.
result F1 scores comparable to BERTBASE with significant speed improvements.
Deep neural networks and decision trees operate on largely separate paradigms; typically, the former performs representation learning with pre-specified architectures, while the latter is characterised by learning hierarchies over pre-specified features with data-driven architectures. We unite the two via adaptive neur…
Coresets are compact representations of data sets such that models trained on a coreset are provably competitive with models trained on the full data set. As such, they have been successfully used to scale up clustering models to massive data sets. While existing approaches generally only allow for multiplicative appro…
ALIEN improves uncertainty estimation of language models by refining entropy-based methods.
problem Overconfidence in uncertainty estimation for language models, especially for difficult inputs.
method ALIEN refines entropy-based uncertainty by aligning it with prediction reliability, using a lightweight uncertainty head.
result ALIEN consistently outperforms strong baselines in detecting incorrect predictions and achieving the lowest calibration error.
We introduce an adaptive output-sensitive Metropolis-Hastings algorithm for probabilistic models expressed as programs, Adaptive Lightweight Metropolis-Hastings (AdLMH). The algorithm extends Lightweight Metropolis-Hastings (LMH) by adjusting the probabilities of proposing random variables for modification to improve c…
Adversarial training (AT) is one of the most effective defenses against adversarial attacks for deep learning models. In this work, we advocate incorporating the hypersphere embedding (HE) mechanism into the AT procedure by regularizing the features onto compact manifolds, which constitutes a lightweight yet effective …
A new perceptual adjustment query for metric learning reduces complexity in high-dimensional data.
problem Metric learning in high-dimensional data with limited human feedback.
method Inverted measurement scheme and two-stage estimator for PAQs.
result Sample complexity guarantees for the two-stage estimator of metric learning from PAQs.
A lightweight framework improves convergence and stability of PINNs for complex PDEs.
problem Training instability and reduced accuracy in PINNs for complex PDEs.
method Adaptive curvature correction using secant information to optimize first-order optimizers.
result Consistent improvements in convergence speed, stability, and accuracy over standard optimizers.
Executing deep neural networks for inference on the server-class or cloud backend based on data generated at the edge of Internet of Things is desirable due primarily to the limited compute power of edge devices and the need to protect the confidentiality of the inference neural networks. However, such a remote inferen…
Optimizes neural architecture search to generate novel lightweight models.
problem Over-reliance on expert knowledge limits NAS to local optima, preventing architectural breakthroughs.
method Casts NAS as an optimization problem, introduces a hierarchical graph-based search space, and uses Bayesian optimization.
result Generates extremely lightweight yet competitive models on six benchmark datasets.
CogScale benchmarks AI architectures for sequential processing.
problem Evaluating AI architectures' ability to process sequential information efficiently.
method 14 scalable synthetic tasks designed to isolate cognitive and memory abilities at different scales.
result Attention mechanisms and modern state-space models consistently maintain high performance as task difficulty scales.
We study the problem of using low computational cost to automate the choices of learners and hyperparameters for an ad-hoc training dataset and error metric, by conducting trials of different configurations on the given training data. We investigate the joint impact of multiple factors on both trial cost and model erro…
LATTE tackles heterogeneous network embedding challenges with layer-stacked attention.
problem Aggregating higher-order indirect relations in heterogeneous networks.
method Layer-stacked ATTention Embedding (LATTE) that decomposes meta relations at each layer.
result LATTE achieves state-of-the-art performance on benchmark datasets.
Interpretable additive models outperform complex DL and hybrid pipelines for air quality forecasting.
problem Accurate forecasting of urban air pollution for public health and policy guidance.
method Investigated lightweight additive models (FBP, NP) vs. deep learning and hybrid pipelines on Beijing PM2.5 and PM10 data.
result Facebook Prophet consistently outperformed NeuralProphet and traditional models, achieving high R2 values. A new method calibrates scientific models by adding randomness to their predictions.
problem Current scientific foundation models lack calibrated uncertainty.
method Stochastic Attention, which randomizes attention weights using multinomial samples.
result Stochastic Attention achieves the strongest native calibration and sharpest prediction intervals.
Kolmogorov-Arnold network improves GW catalog posterior construction.
problem Efficiently constructing posterior distributions for GW catalogs.
method Using the Kolmogorov-Arnold network to create lightweight neural density estimators.
result Kolmogorov-Arnold network achieves superior interpretability and accuracy in posterior construction.
LATM framework uses LLMs to create and reuse tools for efficient problem-solving.
problem Improving problem-solving capabilities of LLMs with external tools.
method Closed-loop framework with two phases: tool making and tool using. LLMs act as both tool makers and users. Resource-intensive model for tool making, lightweight model for tool using.
result LATM framework achieves performance equivalent to using a powerful LLM for both roles but with significantly reduced costs.
Flatland is a simple, lightweight environment for fast prototyping and testing of reinforcement learning agents. It is of lower complexity compared to similar 3D platforms (e.g. DeepMind Lab or VizDoom), but emulates physical properties of the real world, such as continuity, multi-modal partially-observable states with…
ReRe detects anomalies in real-time for time series data.
problem Real-time anomaly detection for time series data requires human intervention or domain knowledge and high computation complexity.
method ReRe uses two lightweight LSTM models to predict and determine anomalies based on historical data and adaptive thresholds.
result ReRe detects anomalies in real-time without requiring human intervention or domain knowledge.
Method transfers feature representation from large to small models using perception coherence.
problem Transfer feature representation from large to small models.
method Defines perception coherence, proposes loss function to minimize.
result Method outperforms or achieves on-par performance compared to strong baseline methods.
FEDS distills LIC model knowledge into a lightweight student for efficient compression.
problem Efficiently compress images with high performance and low resources.
method FEDS combines feature alignment and entropy-based loss for lightweight compression.
result Student model matches teacher's performance while reducing parameters and speeding up.
iSTFTNet speeds up mel-spectrogram vocoders without sacrificing quality.
problem Efficiently converting mel-spectrograms to speech with minimal computation.
method Replaces convolutional layers with iSTFT after frequency dimension reduction.
result Significant reduction in computational cost with comparable quality.
SWIFT improves time series forecasting on edge devices with wavelet decomposition.
problem Efficient time series forecasting on resource-constrained devices.
method Wavelet decomposition, cross-band fusion, and shared linear mapping.
result SWIFT achieves state-of-the-art performance on multiple datasets.
Proposes QEP to mitigate quantization error propagation in layer-wise post-training quantization.
problem Growth of quantization errors across layers degrades performance, especially in low-bit regimes.
method Quantization Error Propagation (QEP) framework that explicitly propagates and compensates for quantization errors.
result QEP-enhanced layer-wise PTQ achieves substantially higher accuracy, especially in low-bit regimes.
GradPower speeds up language model training with minimal code changes.
problem Slower training of large language models.
method Elementwise sign-power transformation applied to gradients.
result Consistently lower terminal loss across various models and datasets.
This work optimizes DNN inference for energy-harvesting devices by compressing and selectively executing neural network exits.
problem Inference delays and energy inefficiency in energy-harvesting devices.
method Developed a power trace-aware and exit-guided network compression algorithm for multi-exit neural networks.
result Superior accuracy and reduced latency compared to state-of-the-art techniques.
A novel framework combines LLMs and RL for financial portfolio optimization.
problem Optimizing financial portfolios using sentiment analysis and market indicators.
method Hierarchical RL structure with base, meta, and super-agents.
result Achieved a 26% annualized return and Sharpe ratio of 1.2.
Unified pipeline detects multi-turn deception using geometric signals.
problem Detecting multi-turn deceptive interactions in LLMs.
method Geometric signaturization via genetic prompt optimization.
result Compact geometric model achieves high recall and F1 scores.
Simple attack bypasses state-of-the-art DNN watermarking.
problem Protecting DNN models from watermark removal attacks.
method Combining imperceptible pattern embedding and spatial-level transformations for a simple yet effective watermark removal.
result Our attack bypasses state-of-the-art watermarking solutions with high success rates.
Study quantifies how LLMs capture higher-order statistical structure using cumulant expansion.
problem Understanding how LLMs internalize statistical structure during next-token prediction.
method Cumulant-expansion framework treating softmax entropy as perturbation around center distribution.
result Cumulants reveal distinct signatures for mathematical vs. general text prompts, quantifying feature-learning dynamics.
This study compresses BERT to make it suitable for low-capability devices.
problem Large Transformer models are resource-intensive.
method Survey and analysis of model compression techniques for BERT.
result Insights into best practices for compressing large-scale Transformer models.