Me12tt tub

Oct 9, 20120 likes197 views

The document evaluates different feature selection methods for bag-of-words approaches to video categorization. It finds that feature selection can improve results by filtering out non-informative terms. Metadata-based features like tags and descriptions generally outperform visual and audio features, but feature selection provides benefits across different feature types. The best performance comes from combining multiple feature types with transformation and selection techniques.

Feature Selection Methods for Bag-
of-(visual)-Words Approaches
Schmiedeke, Kelm and Sikora
Communication Systems Group
Technische Universität Berlin

4 October, 2012

Motivation 2

sports

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Lessons from last year 3

Features derived from metadata (esp. tags)
outperform visual and ASR ones
• Metadata: Naive Bayes (non translated)
• Visual feat.: SVM (avg. pooled histograms)
• ASR transcripts: kNN (JSD)

Uploader mainly contribute to a single category

Schmiedeke: “Feature Selection Methods for BoW Approaches”

This year‘s question 4

Does feature selection improve results achieved
with BoW model?

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Feature Selection/ Transformation 5

Mutual information:

Term Frequency:

PCA (Eigenvalue decomposition):

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Feature Selection 6

Concepts for terms selection:

Top terms for religion: Top terms for politics: Top terms for health:
bibl (0.0897) lunch (0.1200) jama (0.0495)
jesu (0.0797) obama (0.1113) health (0.0378)
god (0.0796) polit (0.0982) report (0.0357)
unleaven(0.0782) grittv (0.0881) harta (0.0227)
eeli (0.0782) flander (0.0861) exceric (0.0211)
davideel(0.0781) laura (0.0855) yoga (0.0203)
ministri(0.0780) economi(0.0747) study (0.0192)

… … …

daytripp (0.0) sonnet (0.0) ilsr (0.0)
adagio (0.0) screenplai (0.0) resystem (0.0)
acustica (0.0) acustica (0.0) acustica (0.0)

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Feature Selection 7

Top-k-Union:

Top terms for religion: Top terms for politics: Top terms for health:
bibl (0.0897) lunch (0.1200) jama (0.0495)
jesu (0.0797) obama (0.1113) health (0.0378)
god (0.0796) polit (0.0982) report (0.0357)
unleaven(0.0782) grittv (0.0881) harta (0.0227)
eeli (0.0782) flander (0.0861) exceric (0.0211)
davideel(0.0781) laura (0.0855) yoga (0.0203)
misistri(0.0780) economi(0.0747) study (0.0192)

… … …

daytripp (0.0) sonnet (0.0) ilsr (0.0)
adagio (0.0) screenplai (0.0) resystem (0.0)
acustica (0.0) acustica (0.0) acustica (0.0)

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Feature Selection 8

Top-k:

Top terms for religion: Top terms for politics: Top terms for health:
bibl (0.0897) lunch (0.1200) jama (0.0495)
jesu (0.0797) obama (0.1113) health (0.0378)
god (0.0796) polit (0.0982) report (0.0357)
unleaven(0.0782) grittv (0.0881) harta (0.0227)
eeli (0.0782) flander (0.0861) exceric (0.0211)
davideel(0.0781) laura (0.0855) yoga (0.0203)
misistri(0.0780) economi(0.0747) study (0.0192)

… … …

daytripp (0.0) sonnet (0.0) ilsr (0.0)
adagio (0.0) screenplai (0.0) resystem (0.0)
acustica (0.0) acustica (0.0) acustica (0.0)

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Feature Selection 9

Union>th:

Top terms for religion: Top terms for politics: Top terms for health:
bibl (0.0897) lunch (0.1200) jama (0.0495)
jesu (0.0797) obama (0.1113) health (0.0378)
god (0.0796) polit (0.0982) report (0.0357)
unleaven(0.0782) grittv (0.0881) harta (0.0227)
eeli (0.0782) flander (0.0861) exceric (0.0211)
davideel(0.0781) laura (0.0855) yoga (0.0203)
misistri(0.0780) economi(0.0747) study (0.0192)

… … …

daytripp (0.0) sonnet (0.0) ilsr (0.0)
adagio (0.0) screenplai (0.0) resystem (0.0)
acustica (0.0) acustica (0.0) acustica (0.0)
0.0002 0.0002 0.0001

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Feature Selection 10

Intersection>Th:

Top terms for religion: Top terms for politics: Top terms for health:
bibl (0.0897) lunch (0.1200) jama (0.0495)
jesu (0.0797) obama (0.1113) health (0.0378)
god (0.0796) polit (0.0982) report (0.0357)
… … …
web appl gossip
python googl interview
xbox teen iphon
big music san
expo tv texa
… … …
daytripp (0.0) sonnet (0.0) ilsr (0.0)
adagio (0.0) screenplai (0.0) resystem (0.0)
acustica (0.0) acustica (0.0) acustica (0.0)
0.0002 0.0002 0.0001

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Official runs 11

Bag of clustered SURF features transformed
using PCA
• Result does not benefit from transformation

official run without FS/FT
mAP 0.2301 0.2309
CA 41.63 % 41.71 %

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Official runs 12

Bag of filtered ASR transcripts terms (Union>Th)
• Result does benefit from selection

official run without FS/FT
mAP 0.1035 0.0522
CA 32.53 % 26.54 %

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Official runs 13

Bag of clustered SURF features filtered using MI
and intersection>th strategy
• Result does slightly benefit from selection

official run without FS/FT
mAP 0.2259 0.2221
CA 40.80 % 40.78 %

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Official runs 14

Bag of filtered terms derived from tags, title and
descriptions (Union>Th)
• Result does benefit from selection

official run without FS/FT
mAP 0.5225 0.4146
CA 58.18 % 55.70 %

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Official runs 15

Bag of clustered SURF features transformed
using PCA and decision fusion using uploader
• Result does benefit from transformation

official run without FS/FT
mAP 0.3304 0.2988
CA 52.14 % 49.19 %

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Conclusion & Future Work 16

FS showed potential for improving the results

Choice of using MI or TF is not critical, both
methods achieve roughly same results
• Metadata (mAP) : MI12004 (0.5277) vs. TF14976 (0.5275)

Investigation in different scaling schemes (NB)

Use of class-independent selection score (MI)

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Backup 17

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Backup 18

Schmiedeke: “Feature Selection Methods for BoW Approaches”

Extracting visual features 19

SURF are extracted from each key frame
• At keypoints and at a regular grid

Vocabulary is built using hierarchical clustering
on SURF features of development set
• 4096/8196 codewords

Term vector for a single video is obtained by bin-
wise pooling of each key frames’ term vector
• avg

Schmiedeke: “Feature Selection Methods for BoW Approaches”

MediaEval 2012: Tagging Task 20

Question: What is the videos’ blip.tv category?
Blip.tv database (cc): ~ 3300 h
• 5288 training videos
• 9550 test videos
Official evaluation measurement is Mean
Average Precision (mAP)
Workshop will be held 4-5 October 2012 in Pisa,
Italy

Schmiedeke: “Feature Selection Methods for BoW Approaches”

The document evaluates different feature selection methods for bag-of-words approaches. It finds that feature selection can improve results achieved with bag-of-words models, depending on the features and selection method used. When applied to clustered SURF features transformed with PCA, filtered ASR transcripts terms, and metadata tags, the feature selection methods led to improved mean average precision and classification accuracy compared to using the features without selection. The choice of mutual information or term frequency for selection was not critical, as both achieved similar results.

Unit 2 boolean algebra and logic gatesAmrutaMehata

��

This document provides an introduction to Boolean algebra, which describes the behavior of digital circuits. It defines key concepts such as binary values, complement/NOT operations, AND and OR operations. It also outlines several important postulates and theorems of Boolean algebra, including identities, commutativity, absorption, De Morgan's theorems, and Shannon's expansion theorem. The document is intended to teach the basic foundations of Boolean algebra used in digital circuit design and logic gate optimization.

An introduction to variable and feature selectionMarco Meoni

��

Image retrieval based on feature selection methodeSAT Publishing House

��

IJRET : International Journal of Research in Engineering and Technology is an international peer reviewed, online journal published by eSAT Publishing House for the enhancement of research in various disciplines of Engineering and Technology. The aim and scope of the journal is to provide an academic medium and an important reference for the advancement and dissemination of research results that support high-level learning, teaching and research in the fields of Engineering and Technology. We bring together Scientists, Academician, Field Engineers, Scholars and Students of related fields of Engineering and Technology.

Exploratory Analysis of Feature Selection Techniques in Medical Image ProcessingAssociation of Scientists, Developers and Faculties

��

This paper analyses features selection method used in medical image processing. How image is selected by using diverse sort of method similarly: screening, scanning and selecting. We discussed on feature selection procedure which is extensively used for data mining and knowledge discovery and it carryout elimination of redundant features, concomitantly retaining the fundamental bigoted information, feature selection implies less data transmission and efficient data mining. It accentuates the need for further research in the field of pattern recognition that can effectively determine the situation with captured portion of human body.

3. introduction to text miningLokesh Ramaswamy

��

This document provides an introduction to text mining, including defining key concepts such as structured vs. unstructured data, why text mining is useful, and some common challenges. It also outlines important text mining techniques like pre-processing text through normalization, tokenization, stemming, and removing stop words to prepare text for analysis. Text mining methods can be used for applications such as sentiment analysis, predicting markets or customer churn.

Using support vector machine with a hybrid feature selection method to the st...lolokikipipi

��

This document discusses using a support vector machine (SVM) with a hybrid feature selection method to predict stock trends. It proposes using F-score filtering followed by a wrapper method called Supported Sequential Forward Search (SSFS) to select optimal features for the SVM. An experiment applies this approach to NASDAQ index data, reducing 30 features to 17 using F_SSFS and achieving a classification accuracy of 81.7% with the SVM, outperforming a backpropagation neural network. The hybrid approach helps address overfitting issues while improving the SVM's prediction performance.

Text miningAli A Jalil

��

This document discusses text mining and provides an outline of the topic. It defines text mining as the analysis of natural language text data and explains why it is useful given the large amount of unstructured data. The document then describes the basic text mining process, which includes steps like filtering, segmentation, stemming, eliminating excessive words, and clustering. Several applications of text mining are mentioned like call centers, anti-spam, and market intelligence. Challenges of text mining like dealing with unstructured data and large collections of documents are also outlined.

Support Vector machineAnandha L Ranganathan

��

Support Vector Machine (SVM) is a supervised machine learning algorithm that can be used for both classification and regression analysis. It works by finding a hyperplane in an N-dimensional space that distinctly classifies the data points. SVM selects the hyperplane that has the largest distance to the nearest training data points of any class, since larger the margin lower the generalization error of the classifier. SVM can efficiently perform nonlinear classification by implicitly mapping their inputs into high-dimensional feature spaces.

Introduction to Text MiningMinha Hwang

��

The class outline covers introduction to unstructured data analysis, word-level analysis using vector space model and TF-IDF, beyond word-level analysis using natural language processing, and a text mining demonstration in R mining Twitter data. The document provides background on text mining, defines what text mining is and its tasks. It discusses features of text data and methods for acquiring texts. It also covers word-level analysis methods like vector space model and TF-IDF, and applications. It discusses limitations of word-level analysis and how natural language processing can help. Finally, it demonstrates Twitter mining in R.

Support Vector Machine without tearsAnkit Sharma

��

This document provides an overview of support vector machines (SVMs), including their basic concepts, formulations, and applications. SVMs are supervised learning models that analyze data, recognize patterns, and are used for classification and regression. The document explains key SVM properties, the concept of finding an optimal hyperplane for classification, soft margin SVMs, dual formulations, kernel methods, and how SVMs can be used for tasks beyond binary classification like regression, anomaly detection, and clustering.

Support Vector Machinesnextlib

��

This document summarizes support vector machines (SVMs), a machine learning technique for classification and regression. SVMs find the optimal separating hyperplane that maximizes the margin between positive and negative examples in the training data. This is achieved by solving a convex optimization problem that minimizes a quadratic function under linear constraints. SVMs can perform non-linear classification by implicitly mapping inputs into a higher-dimensional feature space using kernel functions. They have applications in areas like text categorization due to their ability to handle high-dimensional sparse data.

Support Vector MachineShao-Chuan Wang

��

This document provides an overview of support vector machines (SVMs). It discusses how SVMs can be used to perform classification tasks by finding optimal separating hyperplanes that maximize the margin between different classes. The document outlines how SVMs solve an optimization problem to find these optimal hyperplanes using techniques like Lagrange duality, kernels, and soft margins. It also covers model selection methods like cross-validation and discusses extensions of SVMs to multi-class classification problems.

Feature Selection in Machine LearningUpekha Vandebona

��

Textmining IntroductionDatamining Tools

��

Text mining refers to extracting knowledge from unstructured text data. It is needed because most biological knowledge exists in unstructured research papers, making it difficult for scientists to manually analyze large amounts of text. Challenges include dealing with noisy, unstructured data and complex relationships between concepts. The text mining process involves preprocessing text through steps like tokenization, feature selection, and parsing to extract meaningful features before analysis can be done through classification, clustering, or other techniques. Potential applications are wide-ranging across domains like customer profiling, trend analysis, and web search.

Support Vector Machines for ClassificationPrakash Pimpale

��

In machine learning, support vector machines (SVMs, also support vector networks[1]) are supervised learning models with associated learning algorithms that analyze data and recognize patterns, used for classification and regression analysis. The basic SVM takes a set of input data and predicts, for each given input, which of two possible classes forms the output, making it a non-probabilistic binary linear classifier.

Feature selectionDong Guo

��

This document summarizes a machine learning workshop on feature selection. It discusses typical feature selection methods like single feature evaluation using metrics like mutual information and Gini indexing. It also covers subset selection techniques like sequential forward selection and sequential backward selection. Examples are provided showing how feature selection improves performance for logistic regression on large datasets with more features than samples. The document outlines the workshop agenda and provides details on when and why feature selection is important for machine learning models.

Feature selection concepts and methodsReza Ramezani

��

This document discusses feature selection concepts and methods. It defines features as attributes that determine which class an instance belongs to. Feature selection aims to select a relevant subset of features by removing irrelevant, redundant and unnecessary data. This improves learning accuracy, model performance and interpretability. The document categorizes feature selection algorithms as filter, wrapper or embedded methods based on how they evaluate feature subsets. It also discusses concepts like feature relevance, search strategies, successor generation and evaluation measures used in feature selection algorithms.

A Review on Feature Selection Methods For Classification TasksEditor IJCATR

��

In recent years, application of feature selection methods in medical datasets has greatly increased. The challenging task in feature selection is how to obtain an optimal subset of relevant and non redundant features which will give an optimal solution without increasing the complexity of the modeling task. Thus, there is a need to make practitioners aware of feature selection methods that have been successfully applied in medical data sets and highlight future trends in this area. The findings indicate that most existing feature selection methods depend on univariate ranking that does not take into account interactions between variables, overlook stability of the selection algorithms and the methods that produce good accuracy employ more number of features. However, developing a universal method that achieves the best classification accuracy with fewer features is still an open research area.

An Introduction to Supervised Machine Learning and Pattern Classification: Th...Sebastian Raschka

��

The document provides an introduction to supervised machine learning and pattern classification. It begins with an overview of the speaker's background and research interests. Key concepts covered include definitions of machine learning, examples of machine learning applications, and the differences between supervised, unsupervised, and reinforcement learning. The rest of the document outlines the typical workflow for a supervised learning problem, including data collection and preprocessing, model training and evaluation, and model selection. Common classification algorithms like decision trees, naive Bayes, and support vector machines are briefly explained. The presentation concludes with discussions around choosing the right algorithm and avoiding overfitting.

UiPath Document Understanding - Generative AI and Active learning capabilitiesDianaGray10

��

This session focus on Generative AI features and Active learning modern experience with Document understanding. Topics Covered: Overview of Document Understanding How Generative Annotation works? What is Generative Classification? How to use Generative Extraction activities? What is Generative Validation? How Active learning modern experience accelerate model training? Q/A ❓ If you have any questions or feedback, please refer to the "Women in Automation 2025" dedicated Forum thread. You can find there extra details and updates.

[Webinar] Scaling Made Simple: Getting Started with No-Code Web AppsSafe Software

��

Ready to simplify workflow sharing across your organization without diving into complex coding? With FME Flow Apps, you can build no-code web apps that make your data work harder for you — fast. In this webinar, we’ll show you how to: Build and deploy Workspace Apps to create an intuitive user interface for self-serve data processing and validation. Automate processes using Automation Apps. Learn to create a no-code web app to kick off workflows tailored to your needs, trigger multiple workspaces and external actions, and use conditional filtering within automations to control your workflows. Create a centralized portal with Gallery Apps to share a collection of no-code web apps across your organization. Through real-world examples and practical demos, you’ll learn how to transform your workflows into intuitive, self-serve solutions that empower your team and save you time. We can’t wait to show you what’s possible!

Build with AI on Google Cloud Session #3Margaret Maynard-Reid

��

Bedrock Data Automation (Preview): Simplifying Unstructured Data ProcessingZilliz

��

Bedrock Data Automation (BDA) is a cloud-based service that simplifies the process of extracting valuable insights from unstructured content—such as documents, images, video, and audio. Come learn how BDA leverages generative AI to automate the transformation of multi-modal data into structured formats, enabling developers to build applications and automate complex workflows with greater speed and accuracy.

Temporary Compound microscope slide .pptxSamir Sharma

��

EaseUS Partition Master Crack 2025 + Serial Keykherorpacca127

��

https://ncracked.com/7961-2/ Note: >> Please copy the link and paste it into Google New Tab now Download link EASEUS Partition Master Crack is a professional hard disk partition management tool and system partition optimization software. It is an all-in-one PC and server disk management toolkit for IT professionals, system administrators, technicians, and consultants to provide technical services to customers with unlimited use. EASEUS Partition Master 18.0 Technician Edition Crack interface is clean and tidy, so all options are at your fingertips. Whether you want to resize, move, copy, merge, browse, check, convert partitions, or change their labels, you can do everything with a few clicks. The defragmentation tool is also designed to merge fragmented files and folders and store them in contiguous locations on the hard drive.

DevNexus - Building 10x Development Organizations.pdfJustin Reock

��

Developer Experience is Dead! Long Live Developer Experience! In this keynote-style session, we’ll take a detailed, granular look at the barriers to productivity developers face today and modern approaches for removing them. 10x developers may be a myth, but 10x organizations are very real, as proven by the influential study performed in the 1980s, ‘The Coding War Games.’ Right now, here in early 2025, we seem to be experiencing YAPP (Yet Another Productivity Philosophy), and that philosophy is converging on developer experience. It seems that with every new method, we invent to deliver products, whether physical or virtual, we reinvent productivity philosophies to go alongside them. But which of these approaches works? DORA? SPACE? DevEx? What should we invest in and create urgency behind today so we don’t have the same discussion again in a decade?

Revolutionizing Field Service: How LLMs Are Powering Smarter Knowledge Access...Earley Information Science

��

Revolutionizing Field Service with LLM-Powered Knowledge Management Field service technicians need instant access to accurate repair information, but outdated knowledge systems often create frustrating delays. Large Language Models (LLMs) are changing the game—enhancing knowledge retrieval, streamlining troubleshooting, and reducing technician dependency on senior staff. In this webinar, Seth Earley and industry experts Sanjay Mehta, and Heather Eisenbraun explore how LLMs and Retrieval-Augmented Generation (RAG) are transforming field service operations. Discover how AI-powered knowledge management is improving efficiency, reducing downtime, and elevating service quality. LLMs for Instant Knowledge Retrieval – How AI-driven search dramatically cuts troubleshooting time. Structured Data & AI – Why high-quality, organized knowledge is essential for LLM success. Real-World Implementation – Lessons from deploying LLM-powered knowledge tools in field service. Business Impact – How AI reduces service delays, optimizes workflows, and enhances technician productivity. Empower your field service teams with AI-driven knowledge access. Watch the webinar to see how LLMs are revolutionizing service efficiency.

Dev Dives: Unlock the future of automation with UiPath Agent BuilderUiPathCommunity

��

This webinar will offer you a first look at the powerful capabilities of UiPath Agent Builder, designed to streamline your automation processes and enhance your workflow efficiency. 📕 During the session, you will: - Discover how to build agents with low-code experience, making it accessible for both developers and business users. - Learn how to leverage automations and activities as tools within your agents, enabling them to handle complex and dynamic workflows. - Gain insights into the AI Trust Layer, which provides robust management and monitoring capabilities, ensuring trust and transparency in your automation processes. - See how agents can be deployed and integrated with your existing UiPath cloud and Studio environments. 👨‍🏫 Speaker: Zach Eslami, Sr. Manager, Product Management Director, UiPath ⏩ Register for our upcoming Dev Dives March session: Unleash the power of macOS Automation with UiPath 👉 AMER: https://bit.ly/Dev_Dives_AMER_March 👉 EMEA & APJ:https://bit.ly/Dev_Dives_EMEA_APJ_March This session was streamed live on February 27, 2025, 15:00 GMT. Check out future Dev Dives 2025 sessions at: 🚩 https://bit.ly/Dev_Dives_2025

THE BIG TEN BIOPHARMACEUTICAL MNCs: GLOBAL CAPABILITY CENTERS IN INDIASrivaanchi Nathan

��

This business intelligence report, "The Big Ten Biopharmaceutical MNCs: Global Capability Centers in India", provides an in-depth analysis of the operations and contributions of the Global Capability Centers (GCCs) of ten leading biopharmaceutical multinational corporations in India. The report covers AstraZeneca, Bayer, Bristol Myers Squibb, GlaxoSmithKline (GSK), Novartis, Sanofi, Roche, Pfizer, Novo Nordisk, and Eli Lilly. In this report each company's GCC is profiled with details on location, workforce size, investment, and the strategic roles these centers play in global business operations, research and development, and information technology and digital innovation.

More Related Content

Viewers also liked (12)

Support Vector machineAnandha L Ranganathan

��

Introduction to Text MiningMinha Hwang

��

Support Vector Machine without tearsAnkit Sharma

��

Support Vector Machinesnextlib

��

Support Vector MachineShao-Chuan Wang

��

Feature Selection in Machine LearningUpekha Vandebona

��

Textmining IntroductionDatamining Tools

��

Support Vector Machines for ClassificationPrakash Pimpale

��

Feature selectionDong Guo

��

Feature selection concepts and methodsReza Ramezani

��

A Review on Feature Selection Methods For Classification TasksEditor IJCATR

��

An Introduction to Supervised Machine Learning and Pattern Classification: Th...Sebastian Raschka

��

Support Vector machineAnandha L Ranganathan

��

Introduction to Text MiningMinha Hwang

��

Support Vector Machine without tearsAnkit Sharma

��

Support Vector Machinesnextlib

��

Support Vector MachineShao-Chuan Wang

��

Feature Selection in Machine LearningUpekha Vandebona

��

Textmining IntroductionDatamining Tools

��

Support Vector Machines for ClassificationPrakash Pimpale

��

Feature selectionDong Guo

��

Feature selection concepts and methodsReza Ramezani

��

A Review on Feature Selection Methods For Classification TasksEditor IJCATR

��

An Introduction to Supervised Machine Learning and Pattern Classification: Th...Sebastian Raschka

��

Recently uploaded (20)

UiPath Document Understanding - Generative AI and Active learning capabilitiesDianaGray10

��

[Webinar] Scaling Made Simple: Getting Started with No-Code Web AppsSafe Software

��

Build with AI on Google Cloud Session #3Margaret Maynard-Reid

��

Bedrock Data Automation (Preview): Simplifying Unstructured Data ProcessingZilliz

��

Temporary Compound microscope slide .pptxSamir Sharma

��

EaseUS Partition Master Crack 2025 + Serial Keykherorpacca127

��

DevNexus - Building 10x Development Organizations.pdfJustin Reock

��

Revolutionizing Field Service: How LLMs Are Powering Smarter Knowledge Access...Earley Information Science

��

Dev Dives: Unlock the future of automation with UiPath Agent BuilderUiPathCommunity

��

THE BIG TEN BIOPHARMACEUTICAL MNCs: GLOBAL CAPABILITY CENTERS IN INDIASrivaanchi Nathan

��

Unlocking DevOps Secuirty :Vault & KeylockHusseinMalikMammadli

��

DevOps iş təhlükəsizliyi sizi maraqlandırır? İstər developer, istər təhlükəsizlik mühəndisi, istərsə də DevOps həvəskarı olun, bu tədbir şəbəkələşmək, biliklərinizi bölüşmək və DevSecOps sahəsində ən son təcrübələri öyrənmək üçün mükəmməl fürsətdir! Bu workshopda DevOps infrastrukturlarının təhlükəsizliyini necə artırmaq barədə danışacayıq. DevOps sistemləri qurularkən avtomatlaşdırılmış, yüksək əlçatan və etibarlı olması ilə yanaşı, həm də təhlükəsizlik məsələləri nəzərə alınmalıdır. Bu səbəbdən, DevOps komandolarının təhlükəsizliyə yönəlmiş praktikalara riayət etməsi vacibdir.

Teaching Prompting and Prompt Sharing to End Users.pptxMichael Blumenthal (Microsoft MVP)

��

Understanding Traditional AI with Custom Vision & MuleSoft.pptxshyamraj55

��

Understanding Traditional AI with Custom Vision & MuleSoft.pptx | ### �ݺ�ߣ Deck Description: This presentation features Atul, a Senior Solution Architect at NTT DATA, sharing his journey into traditional AI using Azure's Custom Vision tool. He discusses how AI mimics human thinking and reasoning, differentiates between predictive and generative AI, and demonstrates a real-world use case. The session covers the step-by-step process of creating and training an AI model for image classification and object detection—specifically, an ad display that adapts based on the viewer's gender. Atulavan highlights the ease of implementation without deep software or programming expertise. The presentation concludes with a Q&A session addressing technical and privacy concerns.

Not a Kubernetes fan? The state of PaaS in 2025Anthony Dahanne

��

Kubernetes won the containers orchestration war. But has it made deploying your apps easier? Let's explore some of Kubernetes extensive app developer tooling, but mainly what the PaaS space looks like in 2025; 18 years after Heroku made it popular. Is Heroku still around? What about Cloud Foundry? And what are those new comers (fly.io, railway, porter.sh, etc.) worth? Did the Cloud giants replace them all?

Combining Lexical and Semantic Search with Milvus 2.5Zilliz

��

In short, lexical search is a way to search your documents based on the keywords they contain, in contrast to semantic search, which compares the similarity of embeddings. We’ll be covering: Why, when, and how should you use lexical search What is the BM25 distance metric How exactly does Milvus 2.5 implement lexical search How to build an improved hybrid lexical + semantic search with Milvus 2.5

Caching for Performance Masterclass: Caching StrategiesScyllaDB

��

L01 Introduction to Nanoindentation - What is hardnessRostislavDaniel

��

Computational Photography: How Technology is Changing Way We Capture the WorldHusseinMalikMammadli

��

📸 Computational Photography (Computer Vision/Image): How Technology is Changing the Way We Capture the World Heç düşünmüsünüzmü, müasir smartfonlar və kameralar necə bu qədər gözəl görüntülər yaradır? Bunun sirri Computational Fotoqrafiyasında(Computer Vision/Imaging) gizlidir—şəkilləri çəkmə və emal etmə üsulumuzu təkmilləşdirən, kompüter elmi ilə fotoqrafiyanın inqilabi birləşməsi.

Leadership u automatizaciji: RPA priče iz prakse!UiPathCommunity

��

Dobrodošli na "AI Powered Automation Leadership Talks", online događaj koji okuplja senior lidere i menadžere iz različitih industrija kako bi podelili svoja iskustva, izazove i strategije u oblasti RPA (Robotic Process Automation). Ovaj događaj pruža priliku da zavirite u način razmišljanja ljudi koji donose ključne odluke u automatizaciji i liderstvu. 📕 Kroz panel diskusiju sa tri izuzetna stručnjaka, istražićemo: Kako uspešno započeti i skalirati RPA projekte u organizacijama. Koji su najveći izazovi u implementaciji RPA-a i kako ih prevazići. Na koje načine automatizacija menja radne procese i pomaže timovima da ostvare više. Bez obzira na vaše iskustvo sa UiPath-om ili RPA uopšte, ovaj događaj je osmišljen kako bi bio koristan svima – od menadžera do tehničkih lidera, i svima koji žele da unaprede svoje razumevanje automatizacije. Pridružite nam se i iskoristite ovu jedinstvenu priliku da naučite od onih koji vode automatizaciju u svojim organizacijama. Pripremite svoja pitanja i inspiraciju za sledeće korake u vašoj RPA strategiji!

2025-02-27 Tech & Play_ Fun, UX, and Community.pdfkatalinjordans1

��

UiPath Document Understanding - Generative AI and Active learning capabilitiesDianaGray10

��

[Webinar] Scaling Made Simple: Getting Started with No-Code Web AppsSafe Software

��

Build with AI on Google Cloud Session #3Margaret Maynard-Reid

��

Bedrock Data Automation (Preview): Simplifying Unstructured Data ProcessingZilliz

��

Temporary Compound microscope slide .pptxSamir Sharma

��

EaseUS Partition Master Crack 2025 + Serial Keykherorpacca127

��

DevNexus - Building 10x Development Organizations.pdfJustin Reock

��

Revolutionizing Field Service: How LLMs Are Powering Smarter Knowledge Access...Earley Information Science

��

Dev Dives: Unlock the future of automation with UiPath Agent BuilderUiPathCommunity

��

THE BIG TEN BIOPHARMACEUTICAL MNCs: GLOBAL CAPABILITY CENTERS IN INDIASrivaanchi Nathan

��

Unlocking DevOps Secuirty :Vault & KeylockHusseinMalikMammadli

��

Teaching Prompting and Prompt Sharing to End Users.pptxMichael Blumenthal (Microsoft MVP)

��

Understanding Traditional AI with Custom Vision & MuleSoft.pptxshyamraj55

��

Not a Kubernetes fan? The state of PaaS in 2025Anthony Dahanne

��

Combining Lexical and Semantic Search with Milvus 2.5Zilliz

��

Caching for Performance Masterclass: Caching StrategiesScyllaDB

��

L01 Introduction to Nanoindentation - What is hardnessRostislavDaniel

��

Computational Photography: How Technology is Changing Way We Capture the WorldHusseinMalikMammadli

��

Leadership u automatizaciji: RPA priče iz prakse!UiPathCommunity

��

2025-02-27 Tech & Play_ Fun, UX, and Community.pdfkatalinjordans1

��

Me12tt tub

1. Feature Selection Methods for Bag- of-(visual)-Words Approaches Schmiedeke, Kelm and Sikora Communication Systems Group Technische Universität Berlin 4 October, 2012

2. Motivation 2 sports Schmiedeke: “Feature Selection Methods for BoW Approaches”

3. Lessons from last year 3 Features derived from metadata (esp. tags) outperform visual and ASR ones • Metadata: Naive Bayes (non translated) • Visual feat.: SVM (avg. pooled histograms) • ASR transcripts: kNN (JSD) Uploader mainly contribute to a single category Schmiedeke: “Feature Selection Methods for BoW Approaches”

4. This year‘s question 4 Does feature selection improve results achieved with BoW model? Schmiedeke: “Feature Selection Methods for BoW Approaches”

5. Feature Selection/ Transformation 5 Mutual information: Term Frequency: PCA (Eigenvalue decomposition): Schmiedeke: “Feature Selection Methods for BoW Approaches”

6. Feature Selection 6 Concepts for terms selection: Top terms for religion: Top terms for politics: Top terms for health: bibl (0.0897) lunch (0.1200) jama (0.0495) jesu (0.0797) obama (0.1113) health (0.0378) god (0.0796) polit (0.0982) report (0.0357) unleaven(0.0782) grittv (0.0881) harta (0.0227) eeli (0.0782) flander (0.0861) exceric (0.0211) davideel(0.0781) laura (0.0855) yoga (0.0203) ministri(0.0780) economi(0.0747) study (0.0192) … … … daytripp (0.0) sonnet (0.0) ilsr (0.0) adagio (0.0) screenplai (0.0) resystem (0.0) acustica (0.0) acustica (0.0) acustica (0.0) Schmiedeke: “Feature Selection Methods for BoW Approaches”

7. Feature Selection 7 Top-k-Union: Top terms for religion: Top terms for politics: Top terms for health: bibl (0.0897) lunch (0.1200) jama (0.0495) jesu (0.0797) obama (0.1113) health (0.0378) god (0.0796) polit (0.0982) report (0.0357) unleaven(0.0782) grittv (0.0881) harta (0.0227) eeli (0.0782) flander (0.0861) exceric (0.0211) davideel(0.0781) laura (0.0855) yoga (0.0203) misistri(0.0780) economi(0.0747) study (0.0192) … … … daytripp (0.0) sonnet (0.0) ilsr (0.0) adagio (0.0) screenplai (0.0) resystem (0.0) acustica (0.0) acustica (0.0) acustica (0.0) Schmiedeke: “Feature Selection Methods for BoW Approaches”

8. Feature Selection 8 Top-k: Top terms for religion: Top terms for politics: Top terms for health: bibl (0.0897) lunch (0.1200) jama (0.0495) jesu (0.0797) obama (0.1113) health (0.0378) god (0.0796) polit (0.0982) report (0.0357) unleaven(0.0782) grittv (0.0881) harta (0.0227) eeli (0.0782) flander (0.0861) exceric (0.0211) davideel(0.0781) laura (0.0855) yoga (0.0203) misistri(0.0780) economi(0.0747) study (0.0192) … … … daytripp (0.0) sonnet (0.0) ilsr (0.0) adagio (0.0) screenplai (0.0) resystem (0.0) acustica (0.0) acustica (0.0) acustica (0.0) Schmiedeke: “Feature Selection Methods for BoW Approaches”

9. Feature Selection 9 Union>th: Top terms for religion: Top terms for politics: Top terms for health: bibl (0.0897) lunch (0.1200) jama (0.0495) jesu (0.0797) obama (0.1113) health (0.0378) god (0.0796) polit (0.0982) report (0.0357) unleaven(0.0782) grittv (0.0881) harta (0.0227) eeli (0.0782) flander (0.0861) exceric (0.0211) davideel(0.0781) laura (0.0855) yoga (0.0203) misistri(0.0780) economi(0.0747) study (0.0192) … … … daytripp (0.0) sonnet (0.0) ilsr (0.0) adagio (0.0) screenplai (0.0) resystem (0.0) acustica (0.0) acustica (0.0) acustica (0.0) 0.0002 0.0002 0.0001 Schmiedeke: “Feature Selection Methods for BoW Approaches”

10. Feature Selection 10 Intersection>Th: Top terms for religion: Top terms for politics: Top terms for health: bibl (0.0897) lunch (0.1200) jama (0.0495) jesu (0.0797) obama (0.1113) health (0.0378) god (0.0796) polit (0.0982) report (0.0357) … … … web appl gossip python googl interview xbox teen iphon big music san expo tv texa … … … daytripp (0.0) sonnet (0.0) ilsr (0.0) adagio (0.0) screenplai (0.0) resystem (0.0) acustica (0.0) acustica (0.0) acustica (0.0) 0.0002 0.0002 0.0001 Schmiedeke: “Feature Selection Methods for BoW Approaches”

11. Official runs 11 Bag of clustered SURF features transformed using PCA • Result does not benefit from transformation official run without FS/FT mAP 0.2301 0.2309 CA 41.63 % 41.71 % Schmiedeke: “Feature Selection Methods for BoW Approaches”

12. Official runs 12 Bag of filtered ASR transcripts terms (Union>Th) • Result does benefit from selection official run without FS/FT mAP 0.1035 0.0522 CA 32.53 % 26.54 % Schmiedeke: “Feature Selection Methods for BoW Approaches”

13. Official runs 13 Bag of clustered SURF features filtered using MI and intersection>th strategy • Result does slightly benefit from selection official run without FS/FT mAP 0.2259 0.2221 CA 40.80 % 40.78 % Schmiedeke: “Feature Selection Methods for BoW Approaches”

14. Official runs 14 Bag of filtered terms derived from tags, title and descriptions (Union>Th) • Result does benefit from selection official run without FS/FT mAP 0.5225 0.4146 CA 58.18 % 55.70 % Schmiedeke: “Feature Selection Methods for BoW Approaches”

15. Official runs 15 Bag of clustered SURF features transformed using PCA and decision fusion using uploader • Result does benefit from transformation official run without FS/FT mAP 0.3304 0.2988 CA 52.14 % 49.19 % Schmiedeke: “Feature Selection Methods for BoW Approaches”

16. Conclusion & Future Work 16 FS showed potential for improving the results Choice of using MI or TF is not critical, both methods achieve roughly same results • Metadata (mAP) : MI12004 (0.5277) vs. TF14976 (0.5275) Investigation in different scaling schemes (NB) Use of class-independent selection score (MI) Schmiedeke: “Feature Selection Methods for BoW Approaches”

17. Backup 17 Schmiedeke: “Feature Selection Methods for BoW Approaches”

18. Backup 18 Schmiedeke: “Feature Selection Methods for BoW Approaches”

19. Extracting visual features 19 SURF are extracted from each key frame • At keypoints and at a regular grid Vocabulary is built using hierarchical clustering on SURF features of development set • 4096/8196 codewords Term vector for a single video is obtained by bin- wise pooling of each key frames’ term vector • avg Schmiedeke: “Feature Selection Methods for BoW Approaches”

20. MediaEval 2012: Tagging Task 20 Question: What is the videos’ blip.tv category? Blip.tv database (cc): ~ 3300 h • 5288 training videos • 9550 test videos Official evaluation measurement is Mean Average Precision (mAP) Workshop will be held 4-5 October 2012 in Pisa, Italy Schmiedeke: “Feature Selection Methods for BoW Approaches”

�ݺ�ߣ

Me12tt tub

Recommended

More Related Content

Viewers also liked (12)

Recently uploaded (20)

Me12tt tub