Open Password – Montag, den 28. Februar 2022

# 1035

Technische Informationsbibliothek – #StandWithUkraine – Verein für Medieninformation und Mediendokumentation – Marianne-Englert-Preis – Denise Ruisinger – Elisabeth Schmitz – F.A.Z.-Archiv – Jasmin Sessler – Bayrischer Rundfunk – Marketing für Archive – Till Wolf – ZDF – Design Thinking – China – Crackdown – Credit-repair Agencies – Fintechs – Ant Group – BOC – Pandemic – Caixin – Liu Xinhai – BIIA

Artificial Intelligence Conference on Search, Data and Text Mining, Analytics and Visualization – Data Analytics – Aishwarya Rajput – Semalytix – Bassam Mokbel – Digitization 2.0 – Deep Learning – Breakout Rooms – Christoph Haxel – Playful Knowledge Transfer – Fraunhofer Institute for Algorithm and Scientific Computing – Brain Drain – Information Gap – Vanessa Lage-Rupprecht – Knowledge Graph – Natural Language Processing – Patent Landscape Mapping – Mazahir Bhagat – Canadian Intellectual Property Office – Data Visualization – AILANI – Klaus Kater – DeepSearch 9 – Early Technology Development – Competitive Intelligence Teams – Linus Whitbald – Ethical Dilemmas of AI – Francisco Webber – Cortical.io – Inefficiency Problem – Energy Consumption – Findability of Information – Von Neumann Gap – Stuck in the Average – Privacy – Analogical Computations – Semantic Foldings – Semantic Fingerprinting

1. 
   Technische Informationsbibliothek
   #StandWithUkraine

II.
Verein für Medieninformation und Mediendokumentation
Die erfolgreichen Newcomer 2022

III.

China

Cracking Down on Credit-Repair Agencies as Cheating Grows

IV.
Titel

A Second Look on AI-SDV 2021: An experience of exchanging and learning, new and elaborate ideas in the exponentially growing field of Data Analytics – By Aishwarya Rajput, Semalytix

Technische Informationsbibliothek

#StandWithUkraine!

#StandwithUkraine: Statement of the TIB Director on the War in Ukraine. TIB shows solidarity with Ukraine, offers help to Ukrainian researchers and stops all Cooperation with Russian partners.

In: Twitter

Verein für Medieninformation
und Mediendokumentation

Die erfolgreichen Newcomer 2022

Der Verein für Medieninformation und Mediendokumentation (vfm) zeichnet seit 2012 jährlich Nachwuchsarbeiten von Studierenden oder Absolventen aus den Bereichen Information, Dokumentation, Archiv und Bibliothek aus. Der Marianne-Englert-Preis wird während der vom vfm organisierten hybrid stattfindenden Frühjahrstagung der Medienarchivare am 26. April 2022 verliehen. In diesem Jahr gehen die mit jeweils 500 Euro dotierten Preise an das Duo Denise Ruisinger und Elisabeth Schmitz, an Jasmin Sessler und an Till Wolf. Alle diesjährigen Preisträgerinnen und Preisträger sind Absolvent:innen des Fachbereichs Media der Hochschule Darmstadt und haben Ende 2021 ihr postgraduales und kooperatives Volontariat mit Zertifikat zum „wissenschaftlichen Dokumentar / Information Specialist“ erfolgreich
abgeschlossen.

Denise Ruisinger und Elisabeth Schmitz (F.A.Z.) werden für ihre Arbeit „Zurück zu den Wurzeln – Das F.A.Z.-Archiv zieht um“ ausgezeichnet. Sie zeigen, dass historische Pressearchive kein „totes Papier“ sind, sondern dass mit den richtigen Methoden durchaus der Weg für eine zukunftsorientierte Nutzung geebnet werden kann.

Jasmin Sessler (Bayrischer Rundfunk) erhält den Preis für ihre Arbeit „Entwicklung einer Marketingstrategie für online bereitgestelltes Archivmaterial“. Sie hat es nicht nur geschafft, ein Maßnahmenpaket für einen messbaren Nutzerzuwachs bei Online-Auftritten von Fernseharchivinhalten zu entwickeln, sondern zusätzlich eine transparente Darstellung vom Verhältnis von Wirkung und Aufwand der von ihr präsentierten Maßnahmen zu erarbeiten.

Till Wolf (ZDF) wird für seine Arbeit „Nutzerzentrierte Innovation durch Design Thinking. Neue methodische Ansätze für agiles Arbeiten im ZDF-Archiv“ geehrt. Er stellt aktuelle Erkenntnisse aus dem Projekt-Management vor und entwickelte zugleich einen anschaulichen Leitfaden für agiles Arbeiten in Medienarchiven mit dem Potential, ein neues, pro-aktiveres Selbstverständnis von Medienarchiven zu beschleunigen.

China

Cracking Down on Credit-Repair Agencies
as Cheating Grows

Credit reports on individuals have become big business in China over the past decade as consumer lending has soared, fueled by the growth of fintech giants such as Ant Group. China’s financial authorities have started a crackdown on companies providing credit-repair services who are cheating customers by claiming they can remove unfavorable information such as unpaid debts from creditreports, and charge excessive fees for doing so.

Credit reports on individuals have become big business in China over the past decade as consumer lending has soared, fueled by the growth of fintech giants such as Ant Group Co. Ltd. Outstanding consumer credit stood at 54.9 trillion yuan ($8.52 trillion) at the end of 2021, up from 10.4 trillion yuan at the end of 2012, PBOC data show.But the ongoing Covid-19 pandemic and a series of crackdowns on sectors such as after-school tutoring have disrupted economic activity and impacted the incomes of millions of individuals, pushing up personal loan delinquencies and defaults.

Against this backdrop, demand for credit-repair services has been rising, a source at the PBOC’s Tianjin branch told Caixin. Some credit-repair service providers have told customers they can delete records related to overdue debts from their credit reports, and some have even claimed they can access the central bank’s credit reporting system and change credit information in it, Caixin has learned from sources with knowledge of the matter. Individuals have reportedly been charged as much as 2,000 yuan or even higher for the service. In some cases, credit-repair businesses have taken advantageof their understanding of policies that are unfamiliar to some individuals to apply on their behalf to correct inaccurate credit records and contravene regulations by charging them for doing so, the source at the PBOC’s Tianjin branch said.

Under China’s Regulations on the Administration of the Credit Reporting Industry, when there are errors or omissions in the information collected, stored or provided by credit reporting institutions, an individual has the right to object and have the information corrected. The institution must respond within 20 days and should correct any false or omitted credit information. Agencies are not allowed to charge fees for such amendments. According to current regulations, neither credit reporting institutions nor commercial banks and other information providers have the right to change or delete credit information that is accurate.

More needs to be done to help individuals correct inaccurate credit records, said Liu Xinhai, a researcher at the financial intelligence research center at Peking University. He suggested credit reporting agencies and lenders expand the channels available to individuals to make it easier for them to have their reports corrected, and that consumer rights protection organizations should consultation services.

Source: BIIA

A Second Look on AI-SDV 2021

An experience of exchanging

and learning, new and elaborate ideas

in the exponentially growing field of Data Analytics

By Aishwarya Rajput, Semalytix

See also: Dr. Bassam Mokbel, Erfahrungsbericht AI-SDV 2021 – An den Fronten der Suche, Datenanalyse, Visualisierung und Wissensverarbeitung, in: Open Password, 21. und 31. Januar 2022

Aishwarya Rajput

The second virtual AI-SDV Conference and exhibition took place over Zoom, with more than 90 attendees and 22 online presentations. The Artificial Intelligence Conference on Search, Data and Text Mining, Analytics and Visualization, focuses on these tools for intelligence, for those in the professional information community. AI-SDV 2021 is a conference for people involved in advanced search and data applications, text mining and visualization technology. Individuals and companies that work in this space presented their latest research findings, tech developments and visions for the future.

The event featured around 22 speakers over two days, plus an exhibition to complement the conference programme of learning, networking and exploring technologies that change the way individuals and institutions work. The focus was on Artificial Intelligence (AI), Digitization 2.0, Deep Learning and other topics identified by specialists working in scientific and technical information. The program included online presentations, and a virtual exhibition by companies to showcase their products. To ensure interaction and compensate for the loss of informal conversations during the breaks, breakout rooms were available to exchange ideas with the speakers and exhibitors alike. Participants raised several questions in chats and after talks. The conference was moderated by Dr Christoph Haxel, who ensured a relaxed and enthusiastic atmosphere.

The conference started with a very insightful talk, „Ping Pong – Playful Knowledge Transfer“, by the Fraunhofer Institute for Algorithms and Scientific Computing SCAI. At the turn of this decade, an immense number of people are retiring from the workforce. The so-called brain drain resulting from the absence of the baby boomers that paved much of the way with all the digitization, modern genome-based medicine and such, is foreseen to create a considerable information gap as they take all their knowledge and expertise with them. The risk of a collective loss of precious knowledge organizations face brings an urgent need to digitize it for future generations to use and develop. With this solid assertion, the speaker, Vanessa Lage-Ruprecht, introduced an authoring tool as a knowledge-digitization guide. It lets professionals play “ping pong” without forcing them to deal with basic principles of formal knowledge representations. In a dynamic, iterative process, the concept overview first detailed by users becomes a highly enriched knowledge graph, which is added to a constantly growing Enterprise, Science & Technology Graph.

The rest of the first day consisted of talks that encompassed and showcased ideas surrounding natural language processing, developing and maintaining ontologies and thesauri, patent landscape mapping and the likes. A talk by Mazahir Bhagat described how the Canadian Intellectual Property Office’s (CIPO) IP Analytics Team uses Intellectual Property (IP) data to showcase the innovation undertaken by Canadian institutions and inventors for specific technology areas or industry sectors. Such analysis is published as reports on the Government of Canada’s website of which data visualization is an important feature since it makes the data easily comprehensible and assists in identifying trends, patterns, and outliers within the data. Another talk entailed how AILANI combines semantic modelling, ontologies, linguistics and artificial intelligence (AI) algorithms in a self-refining system that delivers results based on the inter-related meaning of a fact. Klaus Kater of DeepSearch 9, emphasized the necessity for monitoring early technology development, finding licensing opportunities or acquisition targets, and quick access to broad clinical trial information and close surveillance of the competition, for Competitive Intelligence teams in R&D.

On the second day, after a continuation of some of the topics from the day before, namely „Leveraging pre-trained language models for document classification“, by Karakun, the spotlight was also shined on more philosophical/fundamental questions pertaining to the rise of AI. Linus Whitbald stated how new tools are presenting promising opportunities. However, the use of AI also raises new concerns regarding the transparency of the expected usability and reliability. All in all, a day comprising of topics ranging from project management challenges for IP projects to the ethical dilemmas of AI.

_____________________________________________________

AI algorithms need to be optimized for efficiency rather than precision.

_____________________________________________________

My key takeaway from the conference was the presentation by Francisco Webber, from Cortical.io, „Efficiency is the new precision“, dealing with issues of paramount and urgent interest to leading institutions of research and commerce, alike. As the global data sphere, consisting of machine and human data, continues to grow exponentially, while in comparison, the current von Neumann platform performance is stagnating for many years, we are running into an inefficiency problem grossly underrated by all of us that produce this data, but which decreases the productivity significantly in the long run. AI bypasses the need to understand a system when modelling it; however, this convenience comes with extremely high energy consumption, claims the speaker. The complexity of language makes statistical Natural Language Understanding (NLU) models particularly energy-hungry. Since most of our data sphere consists of human data, we face major obstacles, that Webber lists as follows:

1. Findability of Information – the ability of users to identify appropriate information correlating to keywords is getting worse as data grows.
2. Von Neumann Gap – when processors cannot process faster, we need more of them (increased error rates, power consumption, processing delays).
3. Stuck in the Average – when statistical models generate a bias toward the majority, innovation has a hard time.
4. Privacy – if user profiles are created “passively” on the server-side instead of “actively” on the client-side, we lose control.

The current approach to overcoming these limitations is to use larger and larger data sets on more and more processing nodes for training. Instead, one could argue that AI algorithms need to be optimized for efficiency rather than precision. Weber lists out limitations of von Neumann computing and statistical modelling, in favour of Analogical computations and, what neuroscientists name, Semantics Folding, respectively. Wherein Weber argues, statistical modelling should be disqualified as a brute force approach for language applications. When replacing statistical modelling and arithmetic, set theory and geometry would be, in his opinion, better choices, allowing direct processing of words instead of their occurrence counts, similar to what the human brain does with language using only 7 Watts! On this premise, Weber detailed the approaches of Semantic Fingerprinting, which makes representations of concepts stable across language, and Semantic Folding based Machine Learning.

Overall, it was a good experience of exchanging and learning new and elaborate ideas in the exponentially growing field of Data analytics. In 2022, based on attendee feedback, all the concepts of the ICIC and II-SDV-Meetings will be combined in one event.

Open Password

Forum und Nachrichten
für die Informationsbranche
im deutschsprachigen Raum

Neue Ausgaben von Open Password erscheinen dreimal in der Woche.

Wer den E-Mai-Service kostenfrei abonnieren möchte – bitte unter www.password-online.de eintragen.

Die aktuelle Ausgabe von Open Password ist unmittelbar nach ihrem Erscheinen im Web abzurufen. www.password-online.de/archiv. Das gilt auch für alle früher erschienenen Ausgaben.

International Co-operation Partner:

Outsell (London)
Business Industry Information Association/BIIA (Hongkong)

Anzeige

FAQ + Hilfe