📝 Disappearing Frameworks Explained 👋

"A disappearing framework is an application platform that provides a set of tools to develop web applications in a way that does not rely on web-based technologies, like a browser." [gal30b+] 🤖 #HC #SE

🔗 https://arxiv.org/abs/2309.03201v1 #arxiv

📝 Mi-Go: Test Framework Which Uses YouTube as Data Source for Evaluating Speech Recognition Models Like OpenAI's Whisper 🔊🧠

"Mi-Go is a novel, publicly available testing framework aimed at evaluating the performance and adaptability of general-purpose speech recognition machine learning models across diverse real-world scenarios." [gal30b+] 🤖 #SD #LG #SE

🔗 https://arxiv.org/abs/2309.00329v1 #arxiv

Domain-SEO: Die Bedeutung der URL für Google
Die Qualität des Inhalts, die Benutzerfreundlichkeit der Seite und das Backlink-Portfolio sind nur einige der vielen Kriterien, die Google verwendet, um die Platzierung von Websites in den Suchergebnissen zu bestimmen. Für kommerzielle Online-Projekte ist eine hohe Position in den Suchergebnissen essentiell, da die meisten potenziellen Kunden über Suchmaschinen auf die Seite kommen. #Domain #Google #SE

https://bit.ly/3OYBIoK

📝 ASTER: Automatic Speech Recognition System Accessibility Testing for Stutterers 🔊👾

"ASTER generates stuttering speech test inputs by injecting five different types of stuttering into seed utterances by using three different algorithms, and it can enhance the quality of the test cases with a multi-objective optimization-based seed updating algorithm." [gal30b+] 🤖 #SD #AI #SE

🔗 https://arxiv.org/abs/2308.15742v1 #arxiv

📝 Empowering LLM to Use Smartphone for Intelligent Task Automation 👾

"Proposes AutoDroid, a mobile task automation system that can handle arbitrary tasks on any Android application without manual efforts by combining the commonsense knowledge of LLMs and the domain-specific knowledge of apps." [gal30b+] 🤖 #AI #SE

🔗 https://arxiv.org/abs/2308.15272v1 #arxiv

📝 Generative Model for Models: Rapid DNN Customization for Diverse Tasks and Resource Constraints 👾

"Given an edge scenario, NN-Factory can efficiently customize a compact model specialized in the edge task while satisfying the edge resource constraints by searching for the optimal strategy to assemble the modules." [gal30b+] 🤖 #AI #SE

🔗 https://arxiv.org/abs/2308.15003v1 #arxiv

📝 Try with Simpler -- An Evaluation of Improved Principal Component Analysis in Log-Based Anomaly Detection 🧠

"Proposes an unsupervised technique for log anomaly detection called PCA-SAE based on Principal Component Analysis (PCA) and Semantic-Aware Encoder (SAE)." [gal30b+] 🤖 #LG #SE

⚙️ https://github.com/YangLin-George/SemPCA
🔗 https://arxiv.org/abs/2308.12612v1 #arxiv

📝 An Intentional Forgetting-Driven Self-Healing Method for Deep Reinforcement Learning Systems 🧠👾

"DrDrL is a novel self-healing approach of DRL systems that integrates the intentional forgetting mechanism with continual learning to systematically prioritize the adaptation of the key problem-solving skills." [gal30b+] 🤖 #LG #AI #SE

🔗 https://arxiv.org/abs/2308.12445v1 #arxiv

📝 Deploying Deep Reinforcement Learning Systems: A Taxonomy of Challenges 🧠👾

"The training process of deep reinforcement learning is similar to the traditional reinforcement learning process, except that deep reinforcement learning uses the deep neural network instead of the linear function approximation method." [gal30b+] 🤖 #LG #AI #SE

🔗 https://arxiv.org/abs/2308.12438v1 #arxiv

📝 FECoM: A Step Towards Fine-Grained Energy Measurement for Deep Learning 🧠👾

"A framework for fine-grained energy measurement of DL APby static instrumentation of TensorFlow, which is one of the most popular open-source DL frameworks, with a focus on Convolutional Neural Network (CNN) models." [gal30b+] 🤖 #LG #AI #PF #SE

🔗 https://arxiv.org/abs/2308.12264v1 #arxiv

📝 Ensembling Uncertainty Measures to Improve Safety of Black-Box Classifiers 🧠👾

"SPROUT suspects misclassifications by computing uncertainty measures on the inputs and outputs of a black-box classifier, and blocks the propagation of the output of the classifier to the encompassing system." [gal30b+] 🤖 #LG #AI #SE

🔗 https://arxiv.org/abs/2308.12065v1 #arxiv

📝 Learning Representations on Logs for AIOps 📚👾

"Trains a LLM on public and proprietary log data and demonstrate that it outperforms existing models on multiple downstream tasks, making it a valuable addition to the AIOps platform." [gal30b+] 🤖 #CL #AI #SE

⚙️ https://github.com/huggingface/transformers/blob/main/examples/pytorch
🔗 https://arxiv.org/abs/2308.11526v1 #arxiv

📝 MLOps: A Review 🧠

"Machine learning (ML) is a branch of Artificial Intelligence (AI) that uses statistical and optimization methods to teach computers how to perform a task without requiring them to explicitly follow rules." [gal30b+] 🤖 #LG #SE

🔗 https://arxiv.org/abs/2308.10908v1 #arxiv

📝 Characterizing Usability Issue Discussions in OSS Projects 👋

"Presents an empirical study of usability issues in OSS projects using data science notebook and code editor projects as examples to illustrate how the findings can be applied." [gal30b+] 🤖 #HC #SE

🔗 https://arxiv.org/abs/2308.09876v1 #arxiv

📝 Epicure: Distilling Sequence Model Predictions Into Patterns 🧠

"Epicure takes as input a set of predictions from a sequence model and outputs a lattice structure that represents increasingly generalized patterns that subsume the concrete model predictions and are more likely to generalize to novel examples." [gal30b+] 🤖 #LG #SE

🔗 https://arxiv.org/abs/2308.08203v1 #arxiv

📝 SMARLA: A Safety Monitoring Approach for Deep Reinforcement Learning Agents 🧠👾

"SMARLA uses an ensemble of neural network classifiers to predict safety violations from the agent's states using state abstraction to reduce the state space and facilitate the classification." [gal30b+] 🤖 #LG #AI #SE

🔗 https://arxiv.org/abs/2308.02594v1 #arxiv

📝 Evaluating the Robustness of Test Selection Methods for Deep Neural Networks 🧠

"Identifies potential pitfalls of existing test selection methods and show the existence of these pitfalls through a study with 11 test selection methods, five datasets and two model architectures per dataset." [gal30b+] 🤖 #LG #SE

🔗 https://arxiv.org/abs/2308.01314v1 #arxiv

📝 Enhancing Machine Learning Performance with Continuous in-Session Ground Truth Scores: Pilot Study on Objective Skeletal Muscle Pain Intensity Prediction 🧠👾

"Continuous real-time in-session ground truth pain scores and ANS-modulated endodermal activity (EDA) are recorded via a custom pain platform, which includes a wearable device." [gal30b+] 🤖 #LG #AI #SE

🔗 https://arxiv.org/abs/2308.00886v1 #arxiv

📝 CodeBPE: Investigating Subtokenization Options for Large Language Model Pretraining on Source Code 🧠📚

"Experiments with several subtokenization strategies for code: BPE, SentencePiece, and WordPiece on code, and a custom WordPiece subtokenization that is based on code syntax." [gal30b+] 🤖 #LG #CL #SE

⚙️ https://github.com/wasiahmad/PLBART
🔗 https://arxiv.org/abs/2308.00683v1 #arxiv

📝 DiviML: A Module-Based Heuristic for Mapping Neural Networks Onto Heterogeneous Platforms 🧠

"Develops an approach for compiler-level partitioning of deep neural networks (DNNs) onto multiple interconnected hardware devices to enable heterogeneous DNN execution in datacenters." [gal30b+] 🤖 #LG #DC #SE

⚙️ https://github.com/abdelfattah-lab/diviml
🔗 https://arxiv.org/abs/2308.00127v1 #arxiv