A naukri.com initiative
ML News
Arxiv
1d
345
Image Credit: Arxiv
Adaptive State-Space Mamba for Real-Time Sensor Data Anomaly Detection
- State-space modeling has emerged as a powerful paradigm for sequence analysis in various tasks.
- The proposed Adaptive State-Space Mamba (ASSM) framework is designed for real-time sensor data anomaly detection.
- The framework leverages sequential hidden states and introduces an adaptive gating mechanism for efficient and scalable detection.
- Extensive experiments demonstrate superior performance compared to existing baselines.
Read Full Article
20 Likes
Arxiv
1d
361
Image Credit: Arxiv
LeForecast: Enterprise Hybrid Forecast by Time Series Intelligence
- Demand is spiking in industrial fields for multidisciplinary forecasting, where a broad spectrum of sectors needs planning and forecasts to streamline intelligent business management.
- LeForecast is an enterprise intelligence platform tailored for time series tasks, integrating advanced interpretations of time series data and multi-source information.
- It includes a three-pillar modelling engine combining a large foundation model (Le-TSFM), multimodal model, and hybrid model to drive optimization across multiple sectors in enterprise operations.
- Experimental results verify the efficiency of the platform, making it a profound and practical solution for time series forecasting in various industry use cases.
Read Full Article
21 Likes
Arxiv
1d
364
Image Credit: Arxiv
Ignite Forecasting with SPARK: An Efficient Generative Framework for Refining LLMs in Temporal Knowledge Graph Forecasting
- Ignite Forecasting with SPARK: An Efficient Generative Framework for Refining LLMs in Temporal Knowledge Graph Forecasting
- Temporal Knowledge Graph (TKG) forecasting is crucial for predicting future events using historical data.
- SPARK is a Sequence-level Proxy-Adapting framework for Refining LLMs in TKG forecasting.
- SPARK offers a cost-effective, plug-and-play solution for efficient TKG forecasting by utilizing beam search and traditional TKG models.
Read Full Article
21 Likes
Arxiv
1d
368
Image Credit: Arxiv
Adaptive Clipping for Privacy-Preserving Few-Shot Learning: Enhancing Generalization with Limited Data
- Adaptive Clipping for Privacy-Preserving Few-Shot Learning: Enhancing Generalization with Limited Data
- A new approach called Meta-Clip is introduced for enhancing the utility of privacy-preserving few-shot learning methods.
- The Adaptive Clipping method dynamically adjusts clipping thresholds during training to balance data privacy preservation with learning capacity maximization.
- Experiments demonstrate the effectiveness of Adaptive Clipping in minimizing utility degradation and improving generalization performance.
Read Full Article
22 Likes
Arxiv
1d
76
Image Credit: Arxiv
Advancing Spatiotemporal Prediction using Artificial Intelligence: Extending the Framework of Geographically and Temporally Weighted Neural Network (GTWNN) for Differing Geographical and Temporal Contexts
- This paper proposes an extension to the Geographically and Temporally Weighted Neural Network (GTWNN) framework for spatio-temporal prediction.
- The authors formulate a novel semi-analytical approach to solving Geographically and Temporally Weighted Regression (GTWR) and apply it to London crime data.
- The results demonstrate high-accuracy predictive evaluation scores, validating the assumptions and approximations in the approach.
- The study highlights the importance of considering specific geographic and temporal characteristics when selecting modeling strategies for improved accuracy and suitability.
Read Full Article
4 Likes
Arxiv
1d
79
Image Credit: Arxiv
From Individual to Group: Developing a Context-Aware Multi-Criteria Group Recommender System
- Group decision-making is becoming increasingly common in various domains.
- Conventional recommender systems are not effective in group settings due to their limitations.
- A Context-Aware Multi-Criteria Group Recommender System (CA-MCGRS) is developed to address these challenges.
- CA-MCGRS outperforms other approaches in improving group recommendations by integrating context and multi-criteria evaluations.
Read Full Article
4 Likes
Arxiv
1d
83
Image Credit: Arxiv
Combating the Bullwhip Effect in Rival Online Food Delivery Platforms Using Deep Learning
- The wastage of perishable items has led to significant health and economic crises, increasing business uncertainty and fluctuating customer demand.
- Accurate demand forecasting helps stabilize inventory, optimize supplier orders, and reduce waste.
- A Third-Party Logistics (3PL) supply chain model involving restaurants, online food apps, and customers is presented.
- A deep learning-based demand forecasting model using a two-phase Long Short-Term Memory (LSTM) network is proposed to combat the bullwhip effect in online food delivery platforms.
Read Full Article
5 Likes
Arxiv
1d
197
Image Credit: Arxiv
The Cost of Local and Global Fairness in Federated Learning
- With the emerging application of Federated Learning (FL) in finance, hiring, and healthcare, fairness is crucial to prevent disparities across legally protected attributes like race or gender.
- Global fairness addresses the disparity across the entire population, while local fairness focuses on the disparity within each client.
- This paper introduces a framework that investigates the minimum accuracy lost for enforcing specified levels of global and local fairness in multi-class FL settings.
- Experimental results show that the proposed algorithm outperforms the current state of the art in terms of accuracy-fairness tradeoffs, computational costs, and communication costs.
Read Full Article
11 Likes
Arxiv
1d
311
Image Credit: Arxiv
GroundHog: Revolutionizing GLDAS Groundwater Storage Downscaling for Enhanced Recharge Estimation in Bangladesh
- Long-term groundwater level (GWL) measurement is vital for effective policymaking and recharge estimation.
- A Machine Learning model is developed to mitigate data gaps and accurately predict maximum and minimum GWL.
- An Upsampling Model is trained to produce high-resolution GWLs using low-resolution GLDAS data as input.
- The approach successfully upscales GLDAS data, allowing high-resolution recharge estimations and proactive resource management.
Read Full Article
18 Likes
Arxiv
1d
326
Image Credit: Arxiv
Invariant Control Strategies for Active Flow Control using Graph Neural Networks
- Reinforcement learning (RL) has shown potential in learning complex control strategies for active flow control tasks.
- However, RL applications in turbulent flows are computationally challenging and have limited generalization capabilities.
- To address these limitations, this work proposes the use of graph neural networks (GNNs) for active flow control.
- The results demonstrate that GNN-based control policies achieve comparable performance and improved generalization properties.
Read Full Article
19 Likes
Arxiv
1d
129
Image Credit: Arxiv
Data-driven worker activity recognition and picking efficiency estimation in manual strawberry harvesting
- A practical system was developed to calculate the efficiency of pickers in commercial strawberry harvesting.
- Instrumented picking carts were used to record real-time data of harvested fruit weight, geo-location, and cart movement.
- A CNN-LSTM-based deep neural network was trained to classify a picker's activity into 'Pick' and 'NoPick' classes.
- The technology could aid growers in automated worker activity monitoring and harvest optimization, ultimately enhancing overall harvest efficiency.
Read Full Article
7 Likes
Arxiv
1d
212
Image Credit: Arxiv
Harnessing uncertainty when learning through Equilibrium Propagation in neural networks
- Equilibrium Propagation (EP) is a supervised learning algorithm that trains network parameters using local neuronal activity.
- EP avoids data movement, making it suitable for energy-efficient training on neuromorphic systems.
- EP can learn on hardware with physical uncertainties, providing implications for self-learning systems.
- Research shows successful training of deep neural networks using EP in the presence of uncertainties.
Read Full Article
12 Likes
Arxiv
1d
193
Image Credit: Arxiv
Quamba2: A Robust and Scalable Post-training Quantization Framework for Selective State Space Models
- State Space Models (SSMs) are gaining popularity as an alternative to Transformers due to their memory usage and performance benefits.
- Quamba2 is a post-training quantization framework for selective SSMs that enables scaling on various platforms.
- Quamba2 offers bit-width configurations of W8A8, W4A8, and W4A16, catering to different usage scenarios.
- Experimental results show that Quamba2-8B outperforms other SSM quantization methods, offering significant speed-ups and memory reduction with a minimal accuracy drop.
Read Full Article
11 Likes
Arxiv
1d
300
Image Credit: Arxiv
Task Tokens: A Flexible Approach to Adapting Behavior Foundation Models
- Recent advancements in imitation learning have led to transformer-based behavior foundation models (BFMs) that enable multi-modal, human-like control for humanoid agents.
- The introduction of "Task Tokens" provides a method to tailor BFMs to specific tasks while maintaining flexibility.
- Task Tokens leverage the transformer architecture of BFMs to learn a task-specific encoder through reinforcement learning, allowing the incorporation of user-defined priors and balancing reward design.
- Task Tokens demonstrate efficacy in various tasks, including out-of-distribution scenarios, and are compatible with other prompting modalities.
Read Full Article
18 Likes
Arxiv
1d
326
Image Credit: Arxiv
Learning Library Cell Representations in Vector Space
- Lib2Vec is a self-supervised framework for learning vector representations of library cells.
- The framework includes regularity tests, self-supervised learning, and an attention-based model architecture.
- Experimental results show that Lib2Vec captures functional and electrical similarities.
- Lib2Vec improves circuit learning applications, especially with limited labeled data.
Read Full Article
19 Likes
For uninterrupted reading, download the app