DOCSLIB.ORG

Autotext: an End-To-End Autoai Framework for Text

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Autotext: an End-To-End Autoai Framework for Text PRELIMINARY PREPRINT VERSION: DO NOT CITE The AAAI Digital Library will contain the published version some time after the conference. AutoText: An End-to-End AutoAI Framework for Text Arunima Chaudhary, Alayt Issak∗, Kiran Kate, Yannis Katsis, Abel Valente, Dakuo Wang, Alexandre Evfimievski, Sairam Gurajada, Ban Kawas, Cristiano Malossi, Lucian Popa, Tejaswini Pedapati, Horst Samulowitz, Martin Wistuba, Yunyao Li IBMArchitecture Research AI farunima.chaudhary, yannis.katsis, dakuo.wang, sairam.gurajada, [email protected], [email protected], [email protected], [email protected], fkakate, evfimi, bkawas, lpopa, tejaswinip, samulowitz, [email protected] Abstract ()*&+%&' !"#$%&' Building models for natural language processing (NLP) tasks /#*&0#"&'A&0&.$#>"'B'=>3&0'4;"#"7 remains a daunting task for many, requiring significant tech- I*$#,#H&+- nical expertise, efforts, and resources. In this demonstration, 1%2&0&3 !"#$#%$&' we present AutoText, an end-to-end AutoAI framework for 4&)$'5%$% ()*&+#,&"$ text, to lower the barrier of entry in building NLP models. Au- !"#$%&'$( )#%$*+, toText combines state-of-the-art AutoAI optimization tech- )"#*-.-*%&-'/ niques and learning algorithms for NLP tasks into a single 5%$% E&;+%0 extensible framework. Through its simple, yet powerful UI, /+&*+>.&-->+- F+.8#$&.$;+& non-AI experts (e.g., domain experts) can quickly generate A&%+.8 performant NLP models with support to both control (e.g., G&%$;+#H&+- via specifying constraints) and understand learned models. 5&.0%+%$#C& 4+%"-?>+,&+-@ /#*&0#"& 6#789:;%0#$<' !"-*&.$' (-$#,%$>+- D>,*>-#$#>" Introduction =>3&0- /#*&0#"&- Recent AI advances led to rising desire of businesses and organizations to apply AI to solve their problems (Mao Figure 1: AutoText Architecture et al. 2019). However, developing performant AI models re- !"#$%&'()*+',)-. /0 mains a tedious iterative process that includes data prepara- tion, feature engineering, algorithm/model architecture se- mance: AutoText is able to learn performant models that, lection, hyperparameter tuning, training, model evaluation, in many cases, are comparable or outperform state-of-the- and comparison to other models. To lower the barrier of en- art models for the corresponding tasks/datasets. A video try in AI, the research community has looked into automat- demonstration of AutoText can be found online2. ing this process with AutoML (Hutter, Kotthoff, and Van- Related Work. Despite increasing popularity of AutoAI schoren 2019) or AutoAI (Wang et al. 2019, 2020). systems, we are not aware of any with the comprehensive We present AutoText, an extensible end-to-end AutoAI support for text offered by AutoText. Google AutoML Nat- framework, to democratize the development of Natural Lan- ural Language3 does not allow users to select model types, or guage Processing (NLP) models. Its design considerations give insights into the model produced nor into its tuning and include: (a) Usability: Non-experts can create NLP models hyper-parameter optimization (Weidele et al. 2020). Ama- (referred to as pipelines) through a GUI, while expert users 4 1 zon SageMaker Autopilot offers AutoML capabilities that can fully control the process via a Notebook UI . (b) Exten- can be applied on text, but does not provide intuitive visu- sibility: AutoText employs an extensible architecture, where alizations of pipelines and their performance; it also lacks both the components of the pipelines and the pipeline dis- more advanced capabilities for neural architecture search covery (known as optimization) algorithms are modeled as (NAS). Other systems, such as Microsoft AzureML-BERT5, exchangeable modules. (c) Comprehensiveness: AutoText focus on high-performance model building with BERT pre- offers a comprehensive approach, combining various learn- training and fine-tuning, but do not offer simpler but faster ing algorithms (incl. classical machine learning and deep models. H206 is perhaps the closest AutoML framework to learning approaches) and optimization approaches (incl. al- ours; however, it supports a limited range of learning al- gorithm selection, hyperparameter tuning, and neural archi- tecture search) into a single framework. (d) High Perfor- 2https://youtu.be/tujB0BrYlBw 3 ∗ https://cloud.google.com/natural-language/automl Work done at IBM Research; now at the College of Wooster. 4 Copyright c 2021, Association for the Advancement of Artificial https://aws.amazon.com/sagemaker/autopilot/ Intelligence (www.aaai.org). All rights reserved. 5https://azure.microsoft.com/en-us/services/machine-learning 1This demo focuses on AutoText’s GUI-based interaction. 6https://www.h2o.ai/products/h2o-automl gorithms (e.g., RF, GBM, GLM for classical ML and only These include classical ML operators (TfIdfTransformer, MLP for deep learning), a small pipeline search space (only SVM, XGBoost, etc.) and deep learning-based operators ensembles), and limited features for NLP (Word2Vec). Fi- (BiLSTM, CNN, BERT, XLNet, DistilBERT, MLP, etc.). nally, while NAS is becoming an essential part of Au- The library can be easily extended with new operators to toML (He, Zhao, and Chu 2019), few systems focus on NLP; further increase the coverage of supported techniques. moreover, none combine NAS with exploration of alterna- tive pipelines that may include classical ML algorithms. Search Specification. AutoText supports two ways to define the pipeline search space: (1) Declarative Pipeline Compo- The AutoText Framework sition: Leveraging the declarative LALE framework (Bau- dart et al. 2020), AutoText developers can write high-level AutoText comprises a frontend for easy interaction and a specifications of the search space for each pipeline operator, backend for automated model building. We now describe as well as operator composition rules. (2) Neural architec- each of them in detail. ture search (NAS) enables automatic pipeline composition for deep learning models. Frontend Optimizers. AutoText combines into a single framework End users can initiate the building of NLP models by up- multiple optimization techniques presented in the literature, loading a labeled textual dataset and selecting the target col- including combined algorithm selection and hyperparameter umn. AutoText’s backend handles the rest of the process, 7 tuning (CASH) (Thornton et al. 2013) and NAS (Wistuba, including identifying the task type and suitable evaluation Rawat, and Pedapati 2019). This is done by integrating mul- metric, splitting the dataset into train/test, and discovering tiple optimizers, each suited to different scenarios, selecting pipelines. The best performing pipelines are then shown to between them based on heuristics. These include Hyperopt the user to inspect. This includes both a graphical represen- (Bergstra et al. 2015), Hyperband (Li et al. 2017), ADMM tation of each pipeline illustrating its components and a tab- (CASH optimizer supporting black-box constraints, such as ular leaderboard of pipelines sorted by performance. prediction latency or memory footprint) (Liu et al. 2020), Customizations. AutoText allows users to customize TAPAS (Istrate et al. 2019), and NeuRex (Wistuba 2018). through its UI many aspects of the framework, including Pipeline Selection and Model Tuning. Given a labeled the types of learning algorithms explored, the number of dataset and optional user options, AutoText employs the ap- pipelines returned, as well as constraints, such as a time bud- propriate optimizer to iteratively explore different pipeline get for pipeline discovery. More advanced users can also use configurations. During this process, the optimizer considers its Notebook interface to impose additional controls (e.g., the search specification, potential user constraints, and the override the default search space for individual operators). performance of previously explored configurations to deter- mine the next pipeline to explore. This process continues Backend until the user-specified constraints are satisfied or after a pre- The backend of AutoText includes: (1) Operator library: a determined number of iterations. The resulting pipelines are set of operators that form basic components of typical NLP then shown on the frontend for the user to inspect. pipelines, (2) Search specification, defining the pipeline search space, (3) Optimizers: a set of methods to perform Demonstration joint algorithm search and hyperparameter optimization in We will demonstrate AutoText through a variety of datasets, the defined search space, and (4) Pipeline selection and including standard benchmarks, such as MR (Pang and Lee model tuning: the main component that picks the appropriate 2005) and MPQA (Wiebe, Wilson, and Cardie 2005), and optimizer and coordinates the pipeline discovery process. additional data, such as the US Consumer Financial Com- Operator Library. AutoText employs an extensible opera- plaints dataset8. During the demo the audience will be able tor library, with support for three main operator classes: to gain the following insights: (a) experience how Auto- • Data Preprocessors – operators that clean and/or trans- Text allows the quick generation of AI models, (b) under- form data to prepare for later processing. These include stand common components of NLP models and find out tokenization, lemmatization, part-of-speech (POS) tag- which model architectures work best for particular tasks, by ging, normalization, simplification, spell correction, etc. inspecting the discovered pipelines, and (c) explore trade- offs made by different estimators (such as the training
Load more

Recommended publications
  • The Evolution of Ibm Research Looking Back at 50 Years of Scientific Achievements and Innovations
    FEATURES THE EVOLUTION OF IBM RESEARCH LOOKING BACK AT 50 YEARS OF SCIENTIFIC ACHIEVEMENTS AND INNOVATIONS l Chris Sciacca and Christophe Rossel – IBM Research – Zurich, Switzerland – DOI: 10.1051/epn/2014201 By the mid-1950s IBM had established laboratories in New York City and in San Jose, California, with San Jose being the first one apart from headquarters. This provided considerable freedom to the scientists and with its success IBM executives gained the confidence they needed to look beyond the United States for a third lab. The choice wasn’t easy, but Switzerland was eventually selected based on the same blend of talent, skills and academia that IBM uses today — most recently for its decision to open new labs in Ireland, Brazil and Australia. 16 EPN 45/2 Article available at http://www.europhysicsnews.org or http://dx.doi.org/10.1051/epn/2014201 THE evolution OF IBM RESEARCH FEATURES he Computing-Tabulating-Recording Com- sorting and disseminating information was going to pany (C-T-R), the precursor to IBM, was be a big business, requiring investment in research founded on 16 June 1911. It was initially a and development. Tmerger of three manufacturing businesses, He began hiring the country’s top engineers, led which were eventually molded into the $100 billion in- by one of world’s most prolific inventors at the time: novator in technology, science, management and culture James Wares Bryce. Bryce was given the task to in- known as IBM. vent and build the best tabulating, sorting and key- With the success of C-T-R after World War I came punch machines.
    [Show full text]
  • IBM Research AI Residency Program
    IBM Research AI Residency Program The IBM Research™ AI Residency Program is a 13-month program Topics of focus include: that provides opportunity for scientists, engineers, domain experts – Trust in AI (Causal modeling, fairness, explainability, and entrepreneurs to conduct innovative research and development robustness, transparency, AI ethics) on important and emerging topics in Artificial Intelligence (AI). The program aims at creating and investigating novel approaches – Natural Language Processing and Understanding in AI that progress capabilities towards significant technical (Question and answering, reading comprehension, and real-world challenges. AI Residents work closely with IBM language embeddings, dialog, multi-lingual NLP) Research scientists and are expected to fully complete a project – Knowledge and Reasoning (Knowledge/graph embeddings, within the 13-month residency. The results of the project may neuro-symbolic reasoning) include publications in top AI conferences and journals, development of prototypes demonstrating important new AI functionality – AI Automation, Scaling, and Management (Automated data and fielding of working AI systems. science, neural architecture search, AutoML, transfer learning, few-shot/one-shot/meta learning, active learning, AI planning, As part of the selection process, candidates must submit parallel and distributed learning) a 500-word statement of research interest and goals. – AI and Software Engineering (Big code analysis and understanding, software life cycle including modernize, build, debug, test and manage, software synthesis including refactoring and automated programming) – Human-Centered AI (HCI of AI, human-AI collaboration, AI interaction models and modalities, conversational AI, novel AI experiences, visual AI and data visualization) Deadline to apply: January 31, 2021 Earliest start date: June 1, 2021 Duration: 13 months Locations: IBM Thomas J.
    [Show full text]
  • Treatment and Differential Diagnosis Insights for the Physician's
    Treatment and differential diagnosis insights for the physician’s consideration in the moments that matter most The role of medical imaging in global health systems is literally fundamental. Like labs, medical images are used at one point or another in almost every high cost, high value episode of care. Echocardiograms, CT scans, mammograms, and x-rays, for example, “atlas” the body and help chart a course forward for a patient’s care team. Imaging precision has improved as a result of technological advancements and breakthroughs in related medical research. Those advancements also bring with them exponential growth in medical imaging data. The capabilities referenced throughout this document are in the research and development phase and are not available for any use, commercial or non-commercial. Any statements and claims related to the capabilities referenced are aspirational only. There were roughly 800 million multi-slice exams performed in the United States in 2015 alone. Those studies generated approximately 60 billion medical images. At those volumes, each of the roughly 31,000 radiologists in the U.S. would have to view an image every two seconds of every working day for an entire year in order to extract potentially life-saving information from a handful of images hidden in a sea of data. 31K 800MM 60B radiologists exams medical images What’s worse, medical images remain largely disconnected from the rest of the relevant data (lab results, patient-similar cases, medical research) inside medical records (and beyond them), making it difficult for physicians to place medical imaging in the context of patient histories that may unlock clues to previously unconsidered treatment pathways.
    [Show full text]
  • The Impetus to Creativity in Technology
    The Impetus to Creativity in Technology Alan G. Konheim Professor Emeritus Department of Computer Science University of California Santa Barbara, California 93106 [email protected] [email protected] Abstract: We describe the technical developments ensuing from two well-known publications in the 20th century containing significant and seminal results, a paper by Claude Shannon in 1948 and a patent by Horst Feistel in 1971. Near the beginning, Shannon’s paper sets the tone with the statement ``the fundamental problem of communication is that of reproducing at one point either exactly or approximately a message selected *sent+ at another point.‛ Shannon’s Coding Theorem established the relationship between the probability of error and rate measuring the transmission efficiency. Shannon proved the existence of codes achieving optimal performance, but it required forty-five years to exhibit an actual code achieving it. These Shannon optimal-efficient codes are responsible for a wide range of communication technology we enjoy today, from GPS, to the NASA rovers Spirit and Opportunity on Mars, and lastly to worldwide communication over the Internet. The US Patent #3798539A filed by the IBM Corporation in1971 described Horst Feistel’s Block Cipher Cryptographic System, a new paradigm for encryption systems. It was largely a departure from the current technology based on shift-register stream encryption for voice and the many of the electro-mechanical cipher machines introduced nearly fifty years before. Horst’s vision directed to its application to secure the privacy of computer files. Invented at a propitious moment in time and implemented by IBM in automated teller machines for the Lloyds Bank Cashpoint System.
    [Show full text]
  • IBM Cloud Pak for Data
    IBM Cloud Pak for Data IBM Cloud Pak for Data Intelligently automate your data and AI strategy to connect the right data, to the right people, at the right time, from anywhere. Contents Introduction – IBM Cloud Pak for Data: Any data. Any cloud. Anywhere In today’s uncertain environment every organization must become smarter and more responsive to intelligently operate and engage the – The next generation of world, resiliently respond to market changes, flexibly optimize costs, IBM Cloud Pak for Data and innovate. Fueled by data, AI is empowering leading organizations to transform and deliver value. A recent study found that data-driven – Deployment models for organizations are 178% more likely to outperform in terms of revenue IBM Cloud Pak for Data and profitability. – Top use cases for IBM Cloud However, to successfully scale AI throughout your business, you Pak for Data must overcome data complexity. Companies today are struggling to manage and maintain vast amounts of data spanning public, private – Next steps and on-premises clouds. 70 percent of global respondents stated their company is pulling from over 20 different data sources to inform their AI, BI, and analytics systems. In addition, one third cited data complexity and data siloes as top barriers to AI adoption. Further compounding the complexity of these hybrid data landscapes is the fact that the lifespan of that data—the time that it is most relevant and valuable—is drastically shrinking. 1 The solution? An agile and resilient cloud-native platform that enables clients to predict and automate outcomes with trusted data and AI.
    [Show full text]
  • Introduction to the Cell Multiprocessor
    Introduction J. A. Kahle M. N. Day to the Cell H. P. Hofstee C. R. Johns multiprocessor T. R. Maeurer D. Shippy This paper provides an introductory overview of the Cell multiprocessor. Cell represents a revolutionary extension of conventional microprocessor architecture and organization. The paper discusses the history of the project, the program objectives and challenges, the design concept, the architecture and programming models, and the implementation. Introduction: History of the project processors in order to provide the required Initial discussion on the collaborative effort to develop computational density and power efficiency. After Cell began with support from CEOs from the Sony several months of architectural discussion and contract and IBM companies: Sony as a content provider and negotiations, the STI (SCEI–Toshiba–IBM) Design IBM as a leading-edge technology and server company. Center was formally opened in Austin, Texas, on Collaboration was initiated among SCEI (Sony March 9, 2001. The STI Design Center represented Computer Entertainment Incorporated), IBM, for a joint investment in design of about $400,000,000. microprocessor development, and Toshiba, as a Separate joint collaborations were also set in place development and high-volume manufacturing technology for process technology development. partner. This led to high-level architectural discussions A number of key elements were employed to drive the among the three companies during the summer of 2000. success of the Cell multiprocessor design. First, a holistic During a critical meeting in Tokyo, it was determined design approach was used, encompassing processor that traditional architectural organizations would not architecture, hardware implementation, system deliver the computational power that SCEI sought structures, and software programming models.
    [Show full text]
  • APPLIED MACHINE LEARNING for CYBERSECURITY in SPAM FILTERING and MALWARE DETECTION by Mark Sokolov December, 2020
    APPLIED MACHINE LEARNING FOR CYBERSECURITY IN SPAM FILTERING AND MALWARE DETECTION by Mark Sokolov December, 2020 Director of Thesis: Nic Herndon, PhD Major Department: Computer Science Machine learning is one of the fastest-growing fields and its application to cy- bersecurity is increasing. In order to protect people from malicious attacks, several machine learning algorithms have been used to predict the malicious attacks. This research emphasizes two vulnerable areas of cybersecurity that could be easily ex- ploited. First, we show that spam filtering is a well known problem that has been addressed by many authors, yet it still has vulnerabilities. Second, with the increase of malware threats in our world, a lot of companies use AutoAI to help protect their systems. Nonetheless, AutoAI is not perfect, and data scientists can still design better models. In this thesis I show that although there are efficient mechanisms to prevent malicious attacks, there are still vulnerabilities that could be easily exploited. In the visual spoofing experiment, we show that using a classifier trained on data using Latin alphabet, to classify a message with a combination of Latin and Cyrillic let- ters leads to much lower classification accuracy. In Malware prediction experiment, our model has been able to predict malware attacks on Microsoft computers and got higher accuracy than any well known Auto AI. APPLIED MACHINE LEARNING FOR CYBERSECURITY IN SPAM FILTERING AND MALWARE DETECTION A Thesis Presented to The Faculty of the Department of Computer Science East Carolina University In Partial Fulfillment of the Requirements for the Degree Master of Science in Software Engineering by Mark Sokolov December, 2020 Copyright Mark Sokolov, 2020 APPLIED MACHINE LEARNING FOR CYBERSECURITY IN SPAM FILTERING AND MALWARE DETECTION by Mark Sokolov APPROVED BY: DIRECTOR OF THESIS: Nic Herndon, PhD COMMITTEE MEMBER: Rui Wu, PhD COMMITTEE MEMBER: Mark Hills, PhD CHAIR OF THE DEPARTMENT OF COMPUTER SCIENCE: Venkat Gudivada, PhD DEAN OF THE GRADUATE SCHOOL: Paul J.
    [Show full text]
  • IBM Infosphere
    Software Steve Mills Senior Vice President and Group Executive Software Group Software Performance A Decade of Growth Revenue + $3.2B Grew revenue 1.7X and profit 2.9X + $5.6B expanding margins 13 points $18.2B$18.2B $21.4B$21.4B #1 Middleware Market Leader * $12.6B$12.6B Increased Key Branded Middleware 2000 2006 2009 from 38% to 59% of Software revenues Acquired 60+ companies Pre-Tax Income 34% Increased number of development labs Margin globally from 15 to 42 27% 7 pts Margin 2010 Roadmap Performance Segment PTI Growth Model 12% - 15% $8.1B$8.1B 21% 6 pts • Grew PTI to $8B at a 14% CGR Margin • Expanded PTI Margin by 7 points $5.5B$5.5B $2.8B$2.8B ’00–’06’00–’06 ’06–’09’06–’09 Launched high growth initiatives CGRCGR CGRCGR 12%12% 14%14% • Smarter Planet solutions 2000 2006 2009 • Business Analytics & Optimization GAAP View © 2010 International Business Machines Corporation * Source: IBM Market Insights 04/20/10 Software Will Help Deliver IBM’s 2015 Roadmap IBM Roadmap to 2015 Base Growth Future Operating Portfolio Revenue Acquisitions Leverage Mix Growth Initiatives Continue to drive growth and share gain Accelerate shift to higher value middleware Capitalize on market opportunity * business • Middleware opportunity growth of 5% CGR Invest for growth – High growth products growing 2X faster than rest of • Developer population = 33K middleware Extend Global Reach – Growth markets growing 2X faster than major markets • 42 global development labs with skills in 31 – BAO opportunity growth of 7% countries Acquisitions to extend
    [Show full text]
  • IBM Highlights, 1996-1999
    IBM HIGHLIGHTS, 1996 - 1999 Year Page(s) 1996 2 - 7 1997 7 - 13 1998 13- 21 1999 21 - 26 November 2004 1406HE05 2 1996 Business Performance IBM revenue reaches $75.94 billion, an increase of six percent over 1995, and earnings grow by nearly 30 percent to $5.42 billion. There are 240,615 employees and 622,594 stockholders at year end. Speaking in Atlanta to a group of shareholders, analysts and reporters at the corporation’s annual meeting, IBM chairman Louis V. Gerstner, Jr., discusses IBM’s condition, prospects for growth and the importance of network computing to the company’s future. IBM reaches agreement with the United States Department of Justice to terminate within five years all remaining provisions of the Consent Decree first entered into by IBM and the U.S. government in 1956. Organization IBM forms the Network Computer Division in November. The company says it will operate its worldwide services business under a single brand: IBM Global Services. IBM puts its industry-specific business units on a single global general manager. IBM and Tivoli Systems Inc. enter a merger agreement. Tivoli is a leading provider of systems management software and services for distributed client/server networks of personal computers and workstations. IBM’s acquisition of Tivoli extends the company’s strength in host-based systems management to multiplatform distributed systems. IBM and Edmark Corporation, a developer and publisher of consumer and education software, complete a merger in December. IBM acquires The Wilkerson Group, one of the world’s oldest and largest consulting firms dedicated to the pharmaceutical and medical products industry.
    [Show full text]
  • Q3 2020 2 Leadership Spotlight Leadership Scientific Spotlight 1
    Scientific Updates Q3, 2020 1 Leadership Spotlight Consider Scientific Spotlight 1 Scientific Spotlight 2 downloading Scientific Spotlight 3 this file Publication Highlights Contact us This is an interactive file optimized for PDFs viewers, such as Acrobat Reader and iBooks, and in it you’ll find reference content you might want to go back to. ↖ Use the side menu or the arrow keys to navigate this document Scientific Updates Q3 2020 2 Leadership Spotlight Leadership Scientific Spotlight 1 Scientific Spotlight 2 Spotlight Scientific Spotlight 3 Publication Highlights Contact us Gretchen Jackson, MD, PHD, Dr. Jackson is the director of the Center Vice President and Chief Science Officer, for AI, Research and Evaluation (CARE) IBM Watson Health. at IBM Watson Health and an Associate Professor of Surgery, Pediatrics and Biomedical Informatics at the Vanderbilt University Medical Center. She is an internationally recognized informatician and accomplished clinical surgeon with over 25 years of contributions to informatics research, innovations in health information technologies, and surgical science. continues → Scientific Updates Q3 2020 3 Leadership Spotlight It’s no secret that this year has brought many challenges, and one of the most frustrating things about this year has been Scientific Spotlight 1 dealing with significant uncertainty – from quarantine and working from home, to the rippling effects of COVID-19 on every Scientific Spotlight 2 aspect of life. As the Chief Science Officer for IBM® Watson Health®, a mother of school-aged children and a practicing Scientific Spotlight 3 healthcare provider, I have found it difficult to communicate what we know – and don’t know – about the novel severe acute Publication Highlights respiratory coronavirus that appeared in 2019 (SARS-CoV-2) and the COVID-19 disease.
    [Show full text]
  • IBM Highlights, 1985-1989 (PDF, 145KB)
    IBM HIGHLIGHTS, 1985 -1989 Year Page(s) 1985 2 - 7 1986 7 - 13 1987 13 - 18 1988 18 - 24 1989 24 - 30 February 2003 1406HC02 2 1985 Business Performance IBM’s gross income is $50.05 billion, up nine percent from 1984, and its net earnings are $6.55 billion, up 20 percent from the year before. There are 405,535 employees and 798,152 stockholders at year-end. Organization IBM President John F. Akers succeeds John R. Opel as chief executive officer, effective February 1. Mr. Akers also is to head the Corporate Management Board and serve as chairman of its Policy Committee and Business Operations Committee. PC dealer sales, support and operations are transferred from the Entry Systems Division (ESD) to the National Distribution Division, while the marketing function for IBM’s Personal Computer continues to be an ESD responsibility. IBM announces in September a reorganization of its U.S. marketing operations. Under the realignment, to take effect on Jan. 1, 1986, the National Accounts Division, which markets IBM products to the company’s largest customers, and the National Marketing Division, which serves primarily medium-sized and small customer accounts, are reorganized into two geographic marketing divisions: The North-Central Marketing Division and the South-West Marketing Division. The National Distribution Division, which directs IBM’s marketing efforts through Product Centers, value-added remarketers, and authorized dealers, is to merge its distribution channels, personal computer dealer operations and systems supplies field sales forces into a single sales organization. The National Service Division is to realign its field service operations to be symmetrical with the new marketing organizations.
    [Show full text]
  • Innovator Exercise Steps to Perform This Part of the Exercise We Will Be
    Innovator Predict Employee Turnover using IBM Watson Studio Build, train and evaluate Machine Learning models Exercise steps To perform this part of the exercise we will be using a feature in Watson Studio called AutoAI. AutoAI is an automated machine learning tool that helps to quickly get started and work with machine learning models. It helps in cleaning the data, preparing the data, feature engineering, hyper parameter optimization, model experimentation and then finally choosing the right model to deploy for your data. 1. Let’s start by adding an AutoAI experiment asset type to the project. Click on the Add to project button in the Assets tab of your project and click on AutoAI Experiment. 2. Name the model “Predicting Employee Turnover” and click on Associate a Machine Learning service instance to create a new Machine Learning instance on your cloud. Innovator Predict Employee Turnover using IBM Watson Studio Build, train and evaluate Machine Learning models 3. A new tab will open with the details of the Machine Learning instance. Make sure you’re creating a New Machine Learning instance. Innovator Predict Employee Turnover using IBM Watson Studio Build, train and evaluate Machine Learning models 4. Again, make sure that the lite plan is selected. 5. Leave everything as default, scroll down to the bottom and create the Machine Learning instance. Innovator Predict Employee Turnover using IBM Watson Studio Build, train and evaluate Machine Learning models 6. Confirm the creation of the instance. 7. Once the Machine Learning instance is created, you should be back on the Create an AutoAI experiment details screen.
    [Show full text]