The document discusses searching web forums and summarizing forum content at different granularity levels. It proposes a hierarchical model to represent forum structure with threads, posts, and sentences at different levels. It also describes algorithms like OAKS to generate optimal non-overlapping result sets by maximizing quality scores across levels. Evaluation shows the mixed-granularity approach outperforms methods using only posts in terms of perceived relevance of results for queries. The document also discusses enhancing search using authorship information through multi-dimensional random walks to compute author scores.
The document discusses using contextual information from personal data sources to improve information search and retrieval. It describes how people naturally remember past data based on contextual clues like location, time, other people involved. A personal data assistant could index and integrate content and metadata from various sources to enable contextual searches. Challenges include developing unified data models and tools to discover and leverage both explicit and implicit contextual information from personal information sources.
Personal Information Management Systems - EDBT/ICDT'15 TutorialAmélie Marian
The document discusses the challenges of personal information management systems (PIMS) in the past and potential solutions. It notes that personal data used to be stored in fragmented and disconnected ways across devices, applications and services, making it difficult for users to organize, search and control their data. Early PIMS projects from the late 1990s and 2000s tried to address these issues by developing new models and tools for organizing personal data based on concepts like time, tasks, semantics and social networks. However, personal data remains fragmented across many different systems today. The document proposes that a unified PIMS that centrally manages all of a user's information could help overcome these challenges by giving users more control and freedom over their personal data.
This document appears to be a review analysis of some kind. It does not provide many details about what is being analyzed or reviewed. The summary only conveys that the document relates to analyzing a review, but without more context from the document, it is difficult to determine the specific topic or subject being reviewed.
This document is the 2014 Steel Statistical Yearbook published by the World Steel Association. It provides steel production statistics from 2004-2013 for over 70 countries worldwide. The statistics cover crude steel production, steel production by process, steel product production, steel trade, iron ore production, and apparent steel use. Data is presented in tables with country-level detail.
Two Days Training on Advocacy at Lahore 8 - 9 December 2016sultantareen1976
This two-day training document covers advocacy and how to design an effective advocacy campaign. It defines advocacy as organized efforts by citizens to influence policy formulation and implementation by pressuring authorities. It notes that advocacy campaigns should aim to solve specific problems, empower civil society, and promote democracy. The document outlines four key questions to consider when designing a campaign: what change is wanted, who has decision-making power, what must be done to convince the target, and how will the strategy's effectiveness be measured. It provides steps for an advocacy process, starting with properly identifying and analyzing the problem by understanding its causes and consequences.
Hurricane Sandy homeowners struggle to access $billions in Federal, State, local government, manufacturer and utility rebates and incentives. Many return to their real jobs frustrated and defeated.
I presented to a capacity crowd of 250 (would have been over 300 if we had the room) at the Mount Loretto Community Center on Staten Island. They came to learn what rebates and incentives were available for middle class homeowners so they could rebuild their homes following Hurricane Sandy.
I was glad to help them with what they didn’t know, or thought they knew. Unfortunately there was also quite a bit they couldn’t go backwards to fix.
Here are some of the most prevalent issues where greater coordination can help deliver a much better experience for homeowners both after a disaster or just dealing with the stress of home repair.
The document discusses using contextual information from personal data sources to improve information search and retrieval. It describes how people naturally remember past data based on contextual clues like location, time, other people involved. A personal data assistant could index and integrate content and metadata from various sources to enable contextual searches. Challenges include developing unified data models and tools to discover and leverage both explicit and implicit contextual information from personal information sources.
Personal Information Management Systems - EDBT/ICDT'15 TutorialAmélie Marian
The document discusses the challenges of personal information management systems (PIMS) in the past and potential solutions. It notes that personal data used to be stored in fragmented and disconnected ways across devices, applications and services, making it difficult for users to organize, search and control their data. Early PIMS projects from the late 1990s and 2000s tried to address these issues by developing new models and tools for organizing personal data based on concepts like time, tasks, semantics and social networks. However, personal data remains fragmented across many different systems today. The document proposes that a unified PIMS that centrally manages all of a user's information could help overcome these challenges by giving users more control and freedom over their personal data.
This document appears to be a review analysis of some kind. It does not provide many details about what is being analyzed or reviewed. The summary only conveys that the document relates to analyzing a review, but without more context from the document, it is difficult to determine the specific topic or subject being reviewed.
This document is the 2014 Steel Statistical Yearbook published by the World Steel Association. It provides steel production statistics from 2004-2013 for over 70 countries worldwide. The statistics cover crude steel production, steel production by process, steel product production, steel trade, iron ore production, and apparent steel use. Data is presented in tables with country-level detail.
Two Days Training on Advocacy at Lahore 8 - 9 December 2016sultantareen1976
This two-day training document covers advocacy and how to design an effective advocacy campaign. It defines advocacy as organized efforts by citizens to influence policy formulation and implementation by pressuring authorities. It notes that advocacy campaigns should aim to solve specific problems, empower civil society, and promote democracy. The document outlines four key questions to consider when designing a campaign: what change is wanted, who has decision-making power, what must be done to convince the target, and how will the strategy's effectiveness be measured. It provides steps for an advocacy process, starting with properly identifying and analyzing the problem by understanding its causes and consequences.
Hurricane Sandy homeowners struggle to access $billions in Federal, State, local government, manufacturer and utility rebates and incentives. Many return to their real jobs frustrated and defeated.
I presented to a capacity crowd of 250 (would have been over 300 if we had the room) at the Mount Loretto Community Center on Staten Island. They came to learn what rebates and incentives were available for middle class homeowners so they could rebuild their homes following Hurricane Sandy.
I was glad to help them with what they didn’t know, or thought they knew. Unfortunately there was also quite a bit they couldn’t go backwards to fix.
Here are some of the most prevalent issues where greater coordination can help deliver a much better experience for homeowners both after a disaster or just dealing with the stress of home repair.
The document provides an overview of key concepts about how the Earth changes over time. It is divided into four sections that cover the sun and Earth's rotation/revolution and seasons; the moon, water cycle, weather, and natural disasters; the food chain and producers/consumers; and different sources of energy and protecting the Earth. The content is presented through a series of labeled illustrations and captions to teach these concepts to young students.
SUSE Linux Enterprise is recommended by SAP as the operating system for running SAP applications and SAP HANA. It provides improved performance, high availability, increased reliability, and tightened security compared to other operating systems like UNIX. SUSE Linux Enterprise is optimized for SAP with features like built-in high availability components and an extended update cycle. It has a long partnership with SAP, being the first validated open source high availability solution and virtualization platform for SAP.
The document discusses the history and development of artificial intelligence over the past 70 years. It outlines some of the key milestones in AI research from the early work in the 1950s to modern advances in deep learning. While progress has been significant, fully general human-level AI remains an ongoing challenge that researchers continue working to achieve.
Future office multi employer worksite pptBethany Yorio
1. On a multi-employer worksite, OSHA can hold contractors and subcontractors liable for safety violations even if their own employees are not exposed to hazards.
2. The employers who are typically cited are the exposing employer whose employees face the hazard, the creating employer who caused the hazard, the controlling employer with authority over the worksite, and the correcting employer responsible for fixing hazards.
3. It is important for employers to pre-qualify contractors due to the risk of financial liability for injuries to other employers' employees from safety violations or hazards on the worksite.
Dokumen tersebut membahas cara melakukan backup sistem komputer menggunakan perangkat lunak Norton Ghost 11.5 dengan menghasilkan file image dari partisi hard disk yang memuat sistem operasi. Langkah-langkahnya meliputi memasang Norton Ghost dari CD atau flashdisk, memilih partisi hard disk yang akan dibuat file imagenya, menentukan lokasi penyimpanan file image, dan menunggu proses kompresi file image selesai.
This document provides a summary of three articles:
1) It discusses how general counsels are increasingly taking on risk prediction and management roles within companies to help avoid costly litigation and reputational damage. They are leveraging lessons from past events and collaborating with law firms to assess risks.
2) It describes a lawsuit filed against GNC over the death of an Army private who took the workout supplement Jack3d. The suit claims the supplement's stimulant ingredient was responsible. If found liable, it could expand retailer responsibility for product safety beyond manufacturers.
3) It provides a list of blogs and industry publications worth following for thought leadership, PR news and career advice, including blogs focused on reputation management, leadership
Mercuri international business flash february 2013 ii fornightYogesh Bhat
The document provides industry news and information from various sources. It discusses topics such as:
1) KKR CEO Venky Mysore taking charge of Red Chillies Entertainment as CEO.
2) HP CEO Meg Whitman saying the company will evaluate selling small businesses or projects that don't fit its plans.
3) Indian IT companies waiting to capitalize on outsourcing deals from a vulnerable HP.
3) Top performing executives likely to receive disproportionately higher salary hikes than other employees this year.
Inbound marketing overview for scoping callsBrightIdeas.co
Inbound marketing focuses on creating helpful content like blog posts and assets to attract organic traffic and leads. It works for both business-to-business and business-to-consumer companies by establishing the business as an authority. Creating a lot of valuable evergreen content helps search engine optimization and can be repurposed, giving the business a sustainable competitive advantage through increased traffic and leads over time.
The student learned several things about technologies from creating an opening sequence. They used a high-definition camera with an external microphone, learning about tools like white balance and rule of thirds grids to improve shot quality. For editing, the student used Adobe Premiere Pro more extensively than before, learning new elements like the Fast Color Corrector tool to individually tweak lighting and apply hue, saturation, and white balance adjustments to clips.
09 state of the art of the management of advanced and recurrent ovarian cancerONCOcare
This document discusses several clinical trials evaluating treatments for ovarian cancer. Key findings include:
- The addition of bevacizumab to carboplatin and paclitaxel chemotherapy significantly improved progression-free survival compared to chemotherapy alone in the GOG-218 trial. Median PFS was 14.1 months with continued bevacizumab versus 10.3 months for chemotherapy alone.
- Intraperitoneal chemotherapy resulted in improved progression-free and overall survival compared to intravenous chemotherapy alone in several trials. However, intraperitoneal therapy was also associated with higher rates of adverse events.
- Cytoreductive surgery prior to chemotherapy may improve outcomes compared to chemotherapy before interval debulking surgery, though postoperative complications are
Data mining and machine learning techniques like classification and clustering are increasingly being used to extract useful information from large datasets. Data mining helps provide better customer service and aids scientists in hypothesis formation by analyzing patterns in data from various sources like business transactions, sensor networks, and scientific experiments. Classification algorithms such as decision trees can be applied to datasets containing attributes for individuals and a target variable to predict, like credit worthiness, to build a predictive model. Clustering algorithms like K-means group unlabeled data into clusters without a predefined target variable to discover hidden patterns in the data.
Introductory LogicUnit 6 - Assignment 850 pts.I. For each .docxnormanibarber20063
Blossoms Up! is a flower grower and producer in California that must transform its operations to remain competitive. The company faces higher costs and a long-standing drought. The CEO wants to modernize operations through new technology, but employees fear job losses. Two employees have contacted unions, worrying the CEO. The HR department must address challenges in compensation, talent development, safety, hiring, and employee relations to support the company's strategic transformation while managing change.
The document provides instructions for using an essay tagging service. It outlines a 5-step process: 1) Create an account with required information. 2) Complete a 10-minute order form providing instructions, sources, and deadline. 3) Review bids from writers and choose one based on qualifications. 4) Review the completed paper and authorize payment if satisfied. 5) Request revisions to ensure satisfaction, with a refund offered for plagiarized work. The service aims to fully meet customer needs through high-quality, original content.
This document presents a study on fuzzy relational clustering for sentence-level text clustering. The methodology uses a fuzzy clustering algorithm and develops a sentence similarity method using word-to-word and order similarity. The results show the clustering performance on quotation and news article datasets, with the proposed method achieving comparable performance to other clustering algorithms. Future work includes performing hierarchical fuzzy clustering and updating the text preprocessing.
The document provides an overview of key concepts about how the Earth changes over time. It is divided into four sections that cover the sun and Earth's rotation/revolution and seasons; the moon, water cycle, weather, and natural disasters; the food chain and producers/consumers; and different sources of energy and protecting the Earth. The content is presented through a series of labeled illustrations and captions to teach these concepts to young students.
SUSE Linux Enterprise is recommended by SAP as the operating system for running SAP applications and SAP HANA. It provides improved performance, high availability, increased reliability, and tightened security compared to other operating systems like UNIX. SUSE Linux Enterprise is optimized for SAP with features like built-in high availability components and an extended update cycle. It has a long partnership with SAP, being the first validated open source high availability solution and virtualization platform for SAP.
The document discusses the history and development of artificial intelligence over the past 70 years. It outlines some of the key milestones in AI research from the early work in the 1950s to modern advances in deep learning. While progress has been significant, fully general human-level AI remains an ongoing challenge that researchers continue working to achieve.
Future office multi employer worksite pptBethany Yorio
1. On a multi-employer worksite, OSHA can hold contractors and subcontractors liable for safety violations even if their own employees are not exposed to hazards.
2. The employers who are typically cited are the exposing employer whose employees face the hazard, the creating employer who caused the hazard, the controlling employer with authority over the worksite, and the correcting employer responsible for fixing hazards.
3. It is important for employers to pre-qualify contractors due to the risk of financial liability for injuries to other employers' employees from safety violations or hazards on the worksite.
Dokumen tersebut membahas cara melakukan backup sistem komputer menggunakan perangkat lunak Norton Ghost 11.5 dengan menghasilkan file image dari partisi hard disk yang memuat sistem operasi. Langkah-langkahnya meliputi memasang Norton Ghost dari CD atau flashdisk, memilih partisi hard disk yang akan dibuat file imagenya, menentukan lokasi penyimpanan file image, dan menunggu proses kompresi file image selesai.
This document provides a summary of three articles:
1) It discusses how general counsels are increasingly taking on risk prediction and management roles within companies to help avoid costly litigation and reputational damage. They are leveraging lessons from past events and collaborating with law firms to assess risks.
2) It describes a lawsuit filed against GNC over the death of an Army private who took the workout supplement Jack3d. The suit claims the supplement's stimulant ingredient was responsible. If found liable, it could expand retailer responsibility for product safety beyond manufacturers.
3) It provides a list of blogs and industry publications worth following for thought leadership, PR news and career advice, including blogs focused on reputation management, leadership
Mercuri international business flash february 2013 ii fornightYogesh Bhat
The document provides industry news and information from various sources. It discusses topics such as:
1) KKR CEO Venky Mysore taking charge of Red Chillies Entertainment as CEO.
2) HP CEO Meg Whitman saying the company will evaluate selling small businesses or projects that don't fit its plans.
3) Indian IT companies waiting to capitalize on outsourcing deals from a vulnerable HP.
3) Top performing executives likely to receive disproportionately higher salary hikes than other employees this year.
Inbound marketing overview for scoping callsBrightIdeas.co
Inbound marketing focuses on creating helpful content like blog posts and assets to attract organic traffic and leads. It works for both business-to-business and business-to-consumer companies by establishing the business as an authority. Creating a lot of valuable evergreen content helps search engine optimization and can be repurposed, giving the business a sustainable competitive advantage through increased traffic and leads over time.
The student learned several things about technologies from creating an opening sequence. They used a high-definition camera with an external microphone, learning about tools like white balance and rule of thirds grids to improve shot quality. For editing, the student used Adobe Premiere Pro more extensively than before, learning new elements like the Fast Color Corrector tool to individually tweak lighting and apply hue, saturation, and white balance adjustments to clips.
09 state of the art of the management of advanced and recurrent ovarian cancerONCOcare
This document discusses several clinical trials evaluating treatments for ovarian cancer. Key findings include:
- The addition of bevacizumab to carboplatin and paclitaxel chemotherapy significantly improved progression-free survival compared to chemotherapy alone in the GOG-218 trial. Median PFS was 14.1 months with continued bevacizumab versus 10.3 months for chemotherapy alone.
- Intraperitoneal chemotherapy resulted in improved progression-free and overall survival compared to intravenous chemotherapy alone in several trials. However, intraperitoneal therapy was also associated with higher rates of adverse events.
- Cytoreductive surgery prior to chemotherapy may improve outcomes compared to chemotherapy before interval debulking surgery, though postoperative complications are
Data mining and machine learning techniques like classification and clustering are increasingly being used to extract useful information from large datasets. Data mining helps provide better customer service and aids scientists in hypothesis formation by analyzing patterns in data from various sources like business transactions, sensor networks, and scientific experiments. Classification algorithms such as decision trees can be applied to datasets containing attributes for individuals and a target variable to predict, like credit worthiness, to build a predictive model. Clustering algorithms like K-means group unlabeled data into clusters without a predefined target variable to discover hidden patterns in the data.
Introductory LogicUnit 6 - Assignment 850 pts.I. For each .docxnormanibarber20063
Blossoms Up! is a flower grower and producer in California that must transform its operations to remain competitive. The company faces higher costs and a long-standing drought. The CEO wants to modernize operations through new technology, but employees fear job losses. Two employees have contacted unions, worrying the CEO. The HR department must address challenges in compensation, talent development, safety, hiring, and employee relations to support the company's strategic transformation while managing change.
The document provides instructions for using an essay tagging service. It outlines a 5-step process: 1) Create an account with required information. 2) Complete a 10-minute order form providing instructions, sources, and deadline. 3) Review bids from writers and choose one based on qualifications. 4) Review the completed paper and authorize payment if satisfied. 5) Request revisions to ensure satisfaction, with a refund offered for plagiarized work. The service aims to fully meet customer needs through high-quality, original content.
This document presents a study on fuzzy relational clustering for sentence-level text clustering. The methodology uses a fuzzy clustering algorithm and develops a sentence similarity method using word-to-word and order similarity. The results show the clustering performance on quotation and news article datasets, with the proposed method achieving comparable performance to other clustering algorithms. Future work includes performing hierarchical fuzzy clustering and updating the text preprocessing.
Question Answering as Search - the Anserini Pipeline and Other StoriesSujit Pal
In the last couple of years, we have seen enormous breakthroughs in automated Open Domain Restricted Context Question Answering, also known as Reading Comprehension, where the task is to find an answer to a question from a single document or paragraph. A potentially more useful task is to find an answer for a question from a corpus representing an entire body of knowledge, also known as Open Domain Open Context Question Answering.
To do this, we adapted the BERTSerini architecture (Yang, et al., 2019), using it to answer questions about clinical content from our corpus of 5000+ medical textbooks. The BERTSerini pipeline consists of two components -- a BERT model fine-tuned for Question Answering, and an Anserini (Yang, Fang, and Lin, 2017) IR pipeline for Passage Retrieval. Anserini, in turn, consists of pluggable components for different kinds of query expansion and result reranking. Given a question, Anserini retrieves candidate passages, which the BERT model uses to retrieve the answer from. The best answer is determined using a combination of passage retrieval and answer scores.
Evaluating this system using a locally developed dataset of medical passages, questions, and answers, we adapted the BERT Question Answering component to our content using a combination of fine-tuning with third party SQuAD data, and pre-training the model using our medical content. However, when we replaced the canned passages with passages retrieved using the Anserini pipeline, performance dropped significantly, indicating that the relevance of the retrieved passages was a limiting factor.
The presentation will describe the actions taken to improve the relevance of passages returned by the Anserini pipeline.
A Comparative Analysis of Genetic Algorithm Selection TechniquesIRJET Journal
This document compares different selection techniques used in genetic algorithms. It discusses roulette wheel selection, rank selection, tournament selection, elitism, and steady-state selection. Roulette wheel selection chooses parents based on their fitness, with better fitness having more chances to be selected. Rank selection assigns ranks to the population before selection. Tournament selection randomly chooses individuals and selects the fitter one. Elitism selects the most fit individuals as parents. Steady-state selection keeps most of the population intact between generations. The document provides pros and cons of each technique and concludes with an analysis of their effects on genetic algorithm performance and diversity.
Machine learning algorithms can learn through supervised, unsupervised, or reinforcement learning. Supervised learning involves providing labeled examples to learn a function that maps inputs to outputs. Unsupervised learning identifies hidden patterns in unlabeled data. Reinforcement learning involves an agent learning through trial-and-error interactions with a dynamic environment. Machine learning has applications in areas like computer vision, natural language processing, medical diagnosis, and more.
1. For each of the following code segments, use OpenMP pragmas.docxdurantheseldine
1. For each of the following code segments, use OpenMP pragmas to make the loop parallel, or
explain why the code segment is not suitable for parallel execution.
a. for (i = 0; i < (int) sqrt(x); i++) {
a[i] = i + 12;
if (i < 10) b[i] = a[i];
}
b. flag = 0;
for (i = 0; (i < n) \& (!flag); i++) {
a[i] = 2.8 * i;
if (a[i] < b[i]) flag = 1;
}
c. for (i = 0; i < n; i++) {
a[i] = fun(i);
}
d. for (i = 0; i < n; i++) {
a[i] = fun(i);
if (a[i] < b[i]) b[i] = a[i];
}
e. for (i = 0; i < n; i++) {
a[i] = fun(i);
if (a[i] < b[i]) break;
}
f. product = 0;
for (i = 0; i < n; i++) {
product += a[i] * b[i];
}
g. for (i = j; i < 3 * j; i++) {
a[i] = a[i] + a[i-j];
}
h. for (i = j; i < n; i++) {
a[i] = c * a[i-j];
}
2. Suppose a parallel program completes execution on 32 processors in 348 seconds, and it has
been found that this program spends 21 seconds in initialization and cleanup on one processor, and for
the remaining time all 32 processors are active. What is the scaled speedup of this parallel program?
3. Suppose a parallel program executing on 20 processors spends 98% of its time inside parallel
code. What is the scaled speedup of this parallel program?
4. The table below shows the speedups observed for six different parallel programs A, B, C, D,
E, F as the number of processors is increased from 1 through 8.
Processors Speedup
A B C D E F
1 1.00 1.00 1.00 1.00 1.00 1.00
2 1.60 1.92 1.92 1.96 1.74 1.94
3 2.00 2.73 2.78 2.88 2.30 2.82
4 2.29 3.39 3.57 3.67 2.74 3.65
5 2.50 3.91 4.31 4.46 3.09 4.42
6 2.67 4.29 5.00 5.22 3.38 5.15
7 2.80 4.55 5.65 5.93 3.62 5.84
8 2.91 4.71 6.25 6.25 3.81 6.50
Using the Karp-Flatt metric as the basis, choose the statement that best describes the expected speedup
for each program with 16 processors.
I. The speedup achieved on 16 processors will probably be at least 40% higher than the speedup
achieved on eight processors.
II. The speedup achieved on 16 processors will probably be less than 40% higher than the speedup
achieved on eight processors, due to the increase in overhead as processors are added.
III. The speedup achieved on 16 processors will probably be less than 40% higher than the speedup
achieved on eight processors, due to the large serial component of the computation.
5. Let n ≥ f(p) denote the isoefficiency relation of a parallel system and let M(n) denote the
amount of memory required to store a problem of size n. Use the scalability function to rank the
parallel systems shown below from the most scalable to the least scalable:
a. f(p) = Cp, M(n) = n2.
b. f(p) = C√p, M(n) = n2.
c. f(p) = C√plog p, M(n) = n2.
d. f(p) = Cplog p, M(n) = n2.
e. f(p) = Cp, M(n) = n.
f. f(p) = Cp√p, M(n) = n.
g. f(p) = Cp2√p, M(n) = n.
6. Suppose a problem of size 100,000 can be solved in 15 hours on a computer today. Assuming
that the execution time is solely determined by the CPU speed, d.
This document provides an overview and instructions for an online course on action research. It introduces the instructor and outlines the major assignments, which include a data analysis paper, conclusions paper, action research project, and presentation. It describes the 10 units that make up the course and 4 seminars. Students are instructed to complete tasks for the first two units, which include reviewing their action research proposal and beginning data collection through surveys and interviews. Guidelines are provided on collecting and analyzing both survey and interview data for the upcoming analysis paper. The next steps and expectations from both the instructor and students are also reviewed.
The document provides a review of key concepts for a statistics final exam, including how to calculate regression equations and lines, probabilities using normal and binomial distributions, hypothesis testing, and other statistical analyses. It includes examples of problems and questions that may appear on the exam.
Creating AnswerBot with Keras and TensorFlow (TensorBeat)Avkash Chauhan
With the recent advances into neural networks capabilities to process text and audio data we are very close creating a natural human assistant. TensorFlow from Google is one of the most popular neural network library, and using Keras you can simplify TensorFlow usage. TensorFlow brings amazing capabilities into natural language processing (NLP) and using deep learning, we are expecting bots to become even more smarter, closer to human experience. In this technical discussion, we will explore NLP methods in TensorFlow with Keras to create answer bot, ready to answers specific technical questions. You will learn how to use TensorFlow to train an answer bot, with specific technical questions and use various AWS services to deploy answer bot in cloud.
Tips And Tricks for Teaching Math Online 2Fred Feldon
The document provides tips and strategies for teaching math online effectively. It discusses why students take online classes, success and retention rates being equal to or better than traditional classes. Key differences in teaching online include increased flexibility but also a learning curve and more time required. Using a course management system is recommended over building a course from scratch. Strategies for building a community of learners, supplementing the course with original materials, and preventing cheating are also outlined.
Here are step-by-step directions for finding the mean absolute deviation of a data set:
1. Find the mean (average) of the data set. To do this, add up all the values and divide by the number of values.
2. For each value in the original data set, subtract the mean. This will give you the difference between that value and the mean.
3. Take the absolute value of each difference. The absolute value removes negative signs, so all differences will be positive numbers.
4. Find the mean (average) of the absolute differences. To do this, add up all the absolute differences and divide by the number of values.
5. The result is the mean
Measures of Central Tendency, Mean, Median, Mode
Disclaimer: Some parts of the presentation are obtained from various sources. Credit to the rightful owners.
The document provides practice questions and tips for business mathematics exams. It includes 20 sample questions covering topics like ratios, percentages, time/work problems, profit/loss, and series sums. The questions are multiple choice with explanations provided for the answers.
Mathematics in the Modern World - GE3 - Set TheoryFlipped Channel
If you happen to like this powerpoint, you may contact me at flippedchannel@gmail.com
I offer some educational services like:
-powerpoint presentation maker
-grammarian
-content creator
-layout designer
Subscribe to our online platforms:
FlippED Channel (Youtube)
http://bit.ly/FlippEDChannel
LET in the NET (facebook)
http://bit.ly/LETndNET
The document describes a program called AFTERSCHO☺OL's PGPSE programme for social entrepreneurship. It is a free online program open to all with flexible admission. It includes case studies, articles, study materials and business plan preparation to support students in getting 100% placement or becoming entrepreneurs. Branches have opened in various locations and workshops are conducted on social entrepreneurship. Professional courses can also be pursued along with the program.
The document describes a program called AFTERSCHO☺OL's PGPSE programme for social entrepreneurship. It is a free online program open to all with flexible admission. It includes case studies, articles, study materials and business plan preparation to support students in getting 100% placement or becoming entrepreneurs. Branches have opened in various locations and workshops are conducted on social entrepreneurship. Professional courses can also be pursued along with the program.
Personal Information Search and DiscoveryAmélie Marian
The document discusses the challenges of personal information management in the digital age. It describes how personal data is fragmented across many different devices and systems. Effective personal information management systems are needed to integrate this diverse personal data and support tasks like search, recollecting memories, and knowledge discovery. The author proposes a context-aware personal information management system called the Digital Self Project, which uses a w5h data model to index personal data by context dimensions like what, who, where, when, why and how. Preliminary results show the w5h search approach improves search accuracy over traditional text search.
Personalizing Forum Search using Multidimensional Random WalksAmélie Marian
This document describes a method for improving forum search through personalization. It proposes using multidimensional random walks to compute user similarities based on multiple dimensions like co-participation, interactions, topics and profiles. The approach builds a multidimensional heterogeneous graph and executes random walks with weights based on egocentric relations. Key results show this method predicts similar users to answer future questions better than 6 baselines, and can enhance keyword search by re-ranking results based on contributor authority scores from multiple relations.
Corroborating Facts from Affirmative StatementsAmélie Marian
This document discusses approaches for corroborating information from sources even when the information provided is mostly consistent and does not directly contradict. It proposes assigning multiple trust scores to each source to account for sources having different accuracy levels for different facts. An algorithm is presented that incrementally evaluates facts by selecting groups of facts based on entropy and updating source trust scores. Experiments on real restaurant data demonstrate the approach outperforms existing techniques in precision, recall, and accuracy.
Searching data with substance and styleAmélie Marian
The document discusses semi-structured data processing and personal information search. It describes research at Rutgers University on developing tools for searching semi-structured data that unifies content and structure. The research aims to allow queries to contain both structural and content components and return results even if queries are incomplete. It proposes approaches like defining a unified data model, query relaxations to approximate queries, and scoring frameworks to rank unified search results.
10 Benefits an EPCR Software should Bring to EMS Organizations Traumasoft LLC
The benefits of an ePCR solution should extend to the whole EMS organization, not just certain groups of people or certain departments. It should provide more than just a form for entering and a database for storing information. It should also include a workflow of how information is communicated, used and stored across the entire organization.
DECLARATION OF HELSINKI - History and principlesanaghabharat01
This SlideShare presentation provides a comprehensive overview of the Declaration of Helsinki, a foundational document outlining ethical guidelines for conducting medical research involving human subjects.
Co-Chairs, Val J. Lowe, MD, and Cyrus A. Raji, MD, PhD, prepared useful Practice Aids pertaining to Alzheimer’s disease for this CME/AAPA activity titled “Alzheimer’s Disease Case Conference: Gearing Up for the Expanding Role of Neuroradiology in Diagnosis and Treatment.” For the full presentation, downloadable Practice Aids, and complete CME/AAPA information, and to apply for credit, please visit us at https://bit.ly/3PvVY25. CME/AAPA credit will be available until June 28, 2025.
Kosmoderma Academy, a leading institution in the field of dermatology and aesthetics, offers comprehensive courses in cosmetology and trichology. Our specialized courses on PRP (Hair), DR+Growth Factor, GFC, and Qr678 are designed to equip practitioners with advanced skills and knowledge to excel in hair restoration and growth treatments.
Lecture 6 -- Memory 2015.pptlearning occurs when a stimulus (unconditioned st...AyushGadhvi1
learning occurs when a stimulus (unconditioned stimulus) eliciting a response (unconditioned response) • is paired with another stimulus (conditioned stimulus)
8 Surprising Reasons To Meditate 40 Minutes A Day That Can Change Your Life.pptxHolistified Wellness
We’re talking about Vedic Meditation, a form of meditation that has been around for at least 5,000 years. Back then, the people who lived in the Indus Valley, now known as India and Pakistan, practised meditation as a fundamental part of daily life. This knowledge that has given us yoga and Ayurveda, was known as Veda, hence the name Vedic. And though there are some written records, the practice has been passed down verbally from generation to generation.
5-hydroxytryptamine or 5-HT or Serotonin is a neurotransmitter that serves a range of roles in the human body. It is sometimes referred to as the happy chemical since it promotes overall well-being and happiness.
It is mostly found in the brain, intestines, and blood platelets.
5-HT is utilised to transport messages between nerve cells, is known to be involved in smooth muscle contraction, and adds to overall well-being and pleasure, among other benefits. 5-HT regulates the body's sleep-wake cycles and internal clock by acting as a precursor to melatonin.
It is hypothesised to regulate hunger, emotions, motor, cognitive, and autonomic processes.
low birth weight presentation. Low birth weight (LBW) infant is defined as the one whose birth weight is less than 2500g irrespective of their gestational age. Premature birth and low birth weight(LBW) is still a serious problem in newborn. Causing high morbidity and mortality rate worldwide. The nursing care provide to low birth weight babies is crucial in promoting their overall health and development. Through careful assessment, diagnosis,, planning, and evaluation plays a vital role in ensuring these vulnerable infants receive the specialize care they need. In India every third of the infant weight less than 2500g.
Birth period, socioeconomical status, nutritional and intrauterine environment are the factors influencing low birth weight
Know the difference between Endodontics and Orthodontics.Gokuldas Hospital
Your smile is beautiful.
Let’s be honest. Maintaining that beautiful smile is not an easy task. It is more than brushing and flossing. Sometimes, you might encounter dental issues that need special dental care. These issues can range anywhere from misalignment of the jaw to pain in the root of teeth.
Cell Therapy Expansion and Challenges in Autoimmune DiseaseHealth Advances
There is increasing confidence that cell therapies will soon play a role in the treatment of autoimmune disorders, but the extent of this impact remains to be seen. Early readouts on autologous CAR-Ts in lupus are encouraging, but manufacturing and cost limitations are likely to restrict access to highly refractory patients. Allogeneic CAR-Ts have the potential to broaden access to earlier lines of treatment due to their inherent cost benefits, however they will need to demonstrate comparable or improved efficacy to established modalities.
In addition to infrastructure and capacity constraints, CAR-Ts face a very different risk-benefit dynamic in autoimmune compared to oncology, highlighting the need for tolerable therapies with low adverse event risk. CAR-NK and Treg-based therapies are also being developed in certain autoimmune disorders and may demonstrate favorable safety profiles. Several novel non-cell therapies such as bispecific antibodies, nanobodies, and RNAi drugs, may also offer future alternative competitive solutions with variable value propositions.
Widespread adoption of cell therapies will not only require strong efficacy and safety data, but also adapted pricing and access strategies. At oncology-based price points, CAR-Ts are unlikely to achieve broad market access in autoimmune disorders, with eligible patient populations that are potentially orders of magnitude greater than the number of currently addressable cancer patients. Developers have made strides towards reducing cell therapy COGS while improving manufacturing efficiency, but payors will inevitably restrict access until more sustainable pricing is achieved.
Despite these headwinds, industry leaders and investors remain confident that cell therapies are poised to address significant unmet need in patients suffering from autoimmune disorders. However, the extent of this impact on the treatment landscape remains to be seen, as the industry rapidly approaches an inflection point.
- Video recording of this lecture in English language: https://youtu.be/Pt1nA32sdHQ
- Video recording of this lecture in Arabic language: https://youtu.be/uFdc9F0rlP0
- Link to download the book free: https://nephrotube.blogspot.com/p/nephrotube-nephrology-books.html
- Link to NephroTube website: www.NephroTube.com
- Link to NephroTube social media accounts: https://nephrotube.blogspot.com/p/join-nephrotube-on-social-media.html
Promoting Wellbeing - Applied Social Psychology - Psychology SuperNotesPsychoTech Services
A proprietary approach developed by bringing together the best of learning theories from Psychology, design principles from the world of visualization, and pedagogical methods from over a decade of training experience, that enables you to: Learn better, faster!
1. Amélie Marian – Rutgers University09/30/2013
Searching Web Forums
Amélie Marian, Rutgers University
Joint work with Gayatree Ganu
2. Amélie Marian – Rutgers University09/30/2013
2
Forum Popularity and Search
• Forums with most traffic
[http://rankings.big-boards.com]
- BMW
- 50K uniq visitors/day
- 25M Posts
- 0.6M Members
- Filipino Community
- Subaru Impreza Owners
- Rome Total War
- …
- Pakistan Cricket Fan Site
- Prison Talk
- Online Money making
Despite popularity,
forums lack good
search capabilities
3. Amélie Marian – Rutgers University09/30/2013
3
Patient Emotion and stRucture Search
USer tool(PERSEUS) - Outline
Multi-Granularity Search
Challenges
- Unstructured text
- Background information omitted
- Discussion digression
Contributions
Return each results at varying focus
levels, allowing more or less
context. (CIKM 2013)
Egocentric Search
Challenges
- Multiple interpersonal relations
with varying importance
Contributions
Proposed a multidimensional user
similarity measure.
Use authorship for improving
personalized and keyword search.
4. Amélie Marian – Rutgers University09/30/2013
4
Hierarchical Model
• Hierarchy over objects at three searchable levels
– pertinent sentences, larger posts, entire discussions or threads
• Hierarchy
– captures strength of association, containment relationship
• Lower levels for
smaller objects
• Edge represents
containment
• Edge weight of 2
indicates that the text
in child was repeated
in the text of parent
Thread 1 Thread 2
Post 1 Post 2 Post 4Post 3
Sent 1 Sent 2 Sent 3 Sent 4 Sent 5 Sent 6
Dataset
Word 1 Word 2 Word 3 Word 4 Word 1
2
2
2
5. Amélie Marian – Rutgers University09/30/2013
5
Alternate Scoring Functions
Example Textual Results.
Query : hair loss
Top-4 Results
Post1: (A) Aromasin certainly caused my hair loss and the hair started falling 14 days after the
chemo. However, I bought myself a rather fashionable scarf to hide the baldness. I wear it everyday,
even at home. (B) Onc was shocked by my hair loss so I guess it is unusual on Aromasin. I had no
other side effects from Aromasin, no hot flashes, no stomach aches or muscle pains, no headaches or
nausea and none of the chemo brain.
Post2: (C) Probably everyone is sick of the hair loss questions, but I need help with this falling hair. I
had my first cemotherapy on 16th September, so due in one week for the 2nd treatment. (D) Surely
the hair loss can’t be starting this fast..or can it?. I was running my fingers at the nape of my neck
and about five came out in my fingers. Would love to hear from anyone else have AC done
(Doxorubicin and Cyclophosphamide) only as I am not due to have the 3rd drug (whatever that is - 12
weekly sessions) after the 4 sessions of AC. Doctor said that different people have different side
effects, so I wanted to know what you all went through. (E) Have n’t noticed hair loss elsewhere, just
the top hair and mainly at the back of my neck. (F) I thought the hair would start thining out
between 2nd and 3rd treatment, not weeks after the 1st one. I have very curly long ringlets past my
shoulders and am wondering if it would be better to just cut it short or completely shave it off. I am
willing to try anything to make this stop, does anyone have a good recommendation for a shampoo,
vitamins or supplements and (sadly) a good wig shop in downtown LA.
Post3: My suggestion is, don’t focus so much on organic. Things can be organic and very unhealthy. I
believe it when I read that nothing here is truly organic. They’re allowed a certain percentage. I think
5% of the food can not be organic and it still can carry the organic label. What you want is
nonprocessed, traditional foods. Food that comes from a farm or a farmer’s market. Small farmers are
not organic just because it is too much trouble to get the certification. Their produce is probably better
than most of the industrial organic stuff. (G) Sorry Jennifer, chemotherapy and treatment followed
by hair loss is extremely depressing and you cannot prepare enough for falling hair, especially hair
in clumps. (H) I am on femara and hair loss is non-stop, I had full head of thick hair.
tf*idf
Sent (E) (4.742)
Sent (A) (4.711)
Sent (C) (4.696)
Sent (G) (4.689)
BM25
Sent (D) (10.570)
Sent (B) (10.458)
Sent (H) (10.362)
Sent (E) (10.175)
HScore
Post2 (0.131)
Sent (G) (0.093)
Post1 (0.092)
Sent (H) (0.089)
Score tf*idf (t,d) = (1+log(tft,d)) * log(N/dft) * 1/CharLength
6. Amélie Marian – Rutgers University09/30/2013
6
Scoring Multi-Granularity Results
Goal: Unified scoring for objects at multiple granularity levels
– largely varying sizes
– with inherent containment relationship
Hierarchical Scoring Function (HScore)
Score for node i with respect to search term t and having j children:
… if i is a non-leaf node
= 1 … if i is a leaf node containing t
= 0 … if i is a leaf node not containing t
ewij = edge weight between parent i and child j
P(j) = number of parents of j
C(i) = number of children of i
7. Amélie Marian – Rutgers University09/30/2013
7
Effect of Size Weighting
Parameter on HScore
• Parameter controls the intermixing of granularities
0
2
4
6
8
10
12
14
16
18
20
0 0.1 0.2 0.3 0.4 0.5 BM25
Threads
Posts
Sentences
Size parameter
Numberofresults
intop-20list
HScore
8. Amélie Marian – Rutgers University09/30/2013
8
Multi-Granularity Result Generation
Sorted Ordering:
Post3(2.5), Post1(2.1), Post2(2), Sent1(1.6), Sent2(1.5), Sent3(1.4), Sent4(1.3),
Sent6(0.4), Sent5(0.1), Post4(0.1), Thread1(0.1), Thread2(0.1)
For result size k=4, optimizing for the sum of scores:
• Overlap: {Post3, Post1, Post2, Sent1} Sum Score = 8.2 (minus 1.6?)
• Greedy: {Post3, Post1, Post2, Sent6} Sum Score = 7.0
• Best: {Post3, Post2, Sent1, Sent2} Sum Score = 7.6
33% sample queries had overlap amongst at least 3 of top-10 results
Thread 1 Thread 2
Post 1 Post 2 Post 4Post 3
Sent 1 Sent 2 Sent 3 Sent 4 Sent 5 Sent 6
0.1
2.1 2 2.5 0.1
0.1
0.1 0.41.6 1.5 1.4 1.3
9. Amélie Marian – Rutgers University09/30/2013
9
Multi-Granularity Result Generation
Goal: Generating a non-overlapping result set maximizing
“quality”
• Quality = Sum of scores of all results in the set
• Maximal independent set problem (NP Hard)
• Existing Algorithm: Lexicographic All Independent Sets (LAIS)
outputs maximal independent set with polynomial delay in
specific order
10. Amélie Marian – Rutgers University09/30/2013
10
Optimal Algorithm for k-set
(OAKS)
• Fix node ordering by decreasing scores
• Efficient OAKS Algorithm (typically k<<n):
– Start with k-sized first independent set, i.e., greedy
– Branch from nodes preceding kth node of the set, check if
maximal
– Find new k-sized maximal sets, save in priority queue
– Reject sets from priority queue where starting node occurs
after current best set’s kth node
11. Amélie Marian – Rutgers University09/30/2013
11
OAKS
Sorted Ordering:
Post3(2.5), Post1(2.1), Post2(2), Sent1(1.6), Sent2(1.5), Sent3(1.4), Sent4(1.3),
Sent6(0.4), Sent5(0.1), Post4(0.1), Thread1(0.1), Thread2(0.1)
For k=4, Greedy = {Post3, Post1, Post2, Sent6} SumScore=7.0
In the 1st iteration:
{Post3, Post2, Sent1, Sent2} SumScore = 7.6
{Post3 , Post1, Sent3, Sent4} SumScore = 7.3
Branches from nodes before Sent6,
i.e. Sent1, Sent2, Sent3, Sent4
Branch from Sent1, removing all adjacent to Sent1, {Post3, Post2, Sent1}
Maximal on first 4 nodes? YES!
then complete to size k and insert in queue- {Post3, Post2, Sent1, Sent2}
Thread 1 Thread 2
Post 1 Post 2 Post 4Post 3
Sent 1 Sent 2 Sent 3 Sent 4 Sent 5 Sent 6
0.1
2.1 2 2.5 0.1
0.1
0.1 0.41.6 1.5 1.4 1.3
12. Amélie Marian – Rutgers University09/30/2013
12
Evaluating OAKS Algorithm
Comparing OAKS Runtime
Small overhead for practical k (=20)
• Scoring time = 0.96 sec
• OAKS Result set generation time = 0.09 sec
Word
Frequency
Sets Evaluated Run Time (sec)
LAIS OAKS LAIS OAKS
20-30 57.59 8.12 0.78 0.12
30-40 102.07 5.06 7.88 0.01
40-50 158.80 5.88 26.94 0.01
50-60 410.18 6.30 82.20 0.02
60-70 716.40 5.26 77.61 0.01
70-80 896.59 8.30 143.33 0.04
Comparing LAIS and OAKS
– 100 relatively infrequent queries
with corpus frequency in range
20-30, 30-40…
– OAKS is very efficient. Time
required by OAKS depends on k
OAKS improves over
Greedy SumScore in
31% queries @top20
13. Amélie Marian – Rutgers University09/30/2013
13
Dataset and Evaluation Setting
• Data collected from breastcancer.org
– 31K threads, 301K posts, 1.8M unique sentences, 46K keywords
• 18 Sample Queries
– e.g., broccoli, herceptin side effects, emotional meltdown, scarf or
wig, shampoo recommendation …
• Experimental Search Strategies – top20 results
- Mixed-Hierarchy : Optimal mixed granularity result.
- Posts-Hierarchy : Hierarchical scoring of posts only.
- Posts-tf*idf : Existing traditional search.
- Mixed-BM25
14. Amélie Marian – Rutgers University09/30/2013
14
Evaluating Perceived Relevance
Graded Relevance Scale
Exactly relevant answer,
Relevant but too broad,
Relevant but too narrow,
Partially relevant answer,
Not Relevant
Crowd Sourced Relevance
using Mechanical Turk
- Over 7 annotations
- Quality control -Honey pot
questions
- EM algorithm for consensus
Query = shampoo recommendation
= 0.1 = 0.2 = 0.3 = 0.4
Rank = 1 Rel Broad Rel Broad Rel Broad Partial
2 Rel Broad Rel Broad Rel Broad Partial
3 Rel Broad Rel Broad Rel Broad Partial
4 Rel Broad Rel Broad Exactly Rel Rel Broad
5 Rel Broad Rel Broad Exactly Rel Partial
6 Exactly Rel Exactly Rel Rel Narrow Rel Narrow
7 Rel Broad Exactly Rel Rel Narrow Not Rel
8 Rel Broad Rel Broad Not Rel Partial
9 Rel Broad Rel Narrow Rel Broad Partial
10 Exactly Rel Rel Narrow Partial Rel Narrow
11 Rel Broad Rel Broad Exactly Rel Not Rel
12 Rel Broad Rel Broad Exactly Rel Not Rel
13 Rel Broad Exactly Rel Partial Not Rel
14 Not Rel Exactly Rel Rel Narrow Partial
15 Not Rel Exactly Rel Not Rel Rel Broad
16 Not Rel Rel Broad Rel Narrow Not Rel
17 Exactly Rel Rel Broad Exactly Rel Not Rel
18 Exactly Rel Exactly Rel Partial Partial
19 Not Rel Rel Broad Rel Narrow Not Rel
20 Not Rel Exactly Rel Partial Not Rel
Mixed-Hierarchy
16. Amélie Marian – Rutgers University09/30/2013
16
EgoCentric Search
• Previous technique did not take the authorship of posts into
account
• Some forum participants are similar, sharing same topics of
interest or having the same needs, not necessarily at the
same time
– Rank similar author’s posts higher for personalized search
• Some forum participants are experts, prolific and
knowledgeable
– Expert opinions carry more weight in keyword search
• Author score to enhance personalized & keyword search
17. Amélie Marian – Rutgers University09/30/2013
17
Author Score
• Forum participants have several reasons to be linked
• Build a multidimensional heterogeneous graph over authors
incorporating many relations
• But, users assign different importance to different relations
auth 1
Topic 1
auth 2
auth n
Topic 2
Topic t
query 1
query 2
query n
W(a,t) W(q,t) author 1
author 2
author n
author 3
W(a1,a2)
User Profiles:
- Location
- Age
- Cancer stage
- Treatment
- …
-Co-participation
-Explicit References
18. Amélie Marian – Rutgers University09/30/2013
18
Contributions
Critical problem for leveraging authorship for search:
Incorporating multiple user relations with varying importance
learned egocentrically from user behavior
Outline:
• Author score computation using multidimensional graph
• Personalized predictions of user interactions: authors most
likely to provide answers
• Re-ranking results of keyword search using author expertise
19. Amélie Marian – Rutgers University09/30/2013
19
Multi-Dimensional Random
Walks (MRW)
• Random Walks (RW) for finding most influential users
– Pt+1 = M × Pt … till convergence
– M = α(A + D) + (1 − α)E … relation matrix A, D for dangling
nodes, uniform matrix E, α usually set to 0.85
• Rooted RW for node similarity
– Teleport back to root node with probability (1-α)
– Computes similarity of all nodes w.r.t root node
• Multidimensional RW– Heterogeneous Networks:
– Transition matrix computed as A = 1 * A1 + 2 * A2 + ... + n * An
where i i = 1 and all i >= 0
– Egocentric weights -
For root node r : i (r) = j ewAi (r, m)/ Ak j ewAk (r, j)
… m Ai and j Ak
a
b
c
2
3
A =
a b c
a 0 0 0
b 2 0 0
c 0 3 0
D =
a b c
a 0 0 0.33
b 0 0 0.33
c 0 0 0.33
E =
a b c
a .33 .33 .33
b .33 .33 .33
c .33 .33 .33
20. Amélie Marian – Rutgers University09/30/2013
20
Personalized Answer Search
• Link prediction by leveraging user similarities:
– Given participant behavior, find similar users to the user asking question
– Predict who will respond to this question
• Learn similarities from first 90% training threads
• Relations used:
– Topics covered in text, Co-participation in threads,
Signature profiles, Proximity of posts
• MRW similarity compared with baselines:
– Single relations
– PathSim:
• Existing approach for heterogeneous networks
• Predefined paths of fixed length
• No dynamic choice of path
Link prediction enables
suggesting which threads
or which users to follow
21. Amélie Marian – Rutgers University09/30/2013
21
Predicting User Interactions
0
0.1
0.2
0.3
0.4
0.5
10 20 30 40 50 60 70 80 90 100
MAP
Top-K similar participants
MAP for link prediction
Multidimensional RW
has best prediction
performance
22. Amélie Marian – Rutgers University09/30/2013
22
Predicting User Interactions
• Leverage content of the initial post to find users who are
experts on the question
– TopicScore computed as cosine similarity between author’s history and
initial post
– UserScore = β * MRWScore + (1- β) * TopicScore
Neighbors β = 0 β = 0.1 β = 0.2 β = 1
Top 5 0.52 0.64 (8%) 0.61 (4%) 0.59
Top 10 0.31 0.50 (8%) 0.49 (5%) 0.46
Top 15 0.24 0.43 (8%) 0.42 (6%) 0.40
Top 20 0.20 0.39 (6%) 0.39 (7%) 0.37
Purely MRW
Purely topical
expertise
% Improvement over purely MRW
MAP
23. Amélie Marian – Rutgers University09/30/2013
23
0.72
0.73
0.74
0.75
0.76
0.77
0.78
0.79
0.80
0.81
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
MAP@10
Tradeoff Parameter ω
IR Score λ=0.1
IR Score λ=0.2
Enhanced Keyword Search
• Non-rooted RW to find most influential expert users
• Re-rank top-k results of IR scoring using author scores
• Final score of post = ω*IR_score λ + (1- ω)*Authority_score
– Posts only, tf*idf scoring with size parameter
Re-ranking search
results with author
score yields higher
MAP relevance
4% improvement
5%
24. Amélie Marian – Rutgers University09/30/2013
24
Patient Emotion and stRucture Search
USer tool(PERSEUS) - Conclusions
• Designed hierarchical model and score that allows generating
search results at several granularities of web forum objects.
• Proposed OAKS algorithm for best non-overlapping result.
• Conducted extensive user studies, show that mixed collection of
granularities yields better relevance than post-only results.
• Combined multiple relations linking users for computing similarities
• Enhanced search results using multidimensional author similarity
• Future Directions:
– Multi-granular search on web pages, blogs, emails. Dynamic focus level
selection.
– Search in and out of context over dialogue, interviews, Q&A.
– Optimal result set selection for targeted advertising, result diversification
– Time sensitive recommendations – Changing friendships, progressive
search needs.
Large amount of unstructured textBackground information is often omittedDigressionTime sensitivity and repetitionsLacking good search capabilities
Alpha = 0.2@10 MAP 31%@20 MAP 34%
Our multidimensional RW approach significantly improves over the single thread co-participation relation by 10% for k = 10 neighbors and 21% for k = 100