Himanshu Joshi -.

Alakh Verma, CEO of Bcentriqe.ai, delivered a compelling and thoughtful talk at the UN forum in Vienna, Austria, on the relevance of AI for developing countries. His address focused on the transformative potential of artificial intelligence in accelerating development and addressing global challenges. Here’s a summary of key points from his speech: AI as a Catalyst for Development: Verma emphasized that AI has the power to leapfrog traditional developmental stages, offering innovative solutions to complex problems such as healthcare, education, and infrastructure. He argued that with the right strategies, developing countries could harness AI to drive significant economic and social progress. Equitable Access to AI: He underscored the importance of ensuring equitable access to AI technologies. Verma pointed out that without inclusive access, there is a risk of widening the digital divide. He advocated for international cooperation to provide resources, training, and infrastructure to developing regions, enabling them to benefit from AI advancements. Ethical AI Implementation: Addressing the ethical dimensions of AI, Verma highlighted the need for responsible and transparent AI practices. He discussed the importance of creating frameworks that prioritize fairness, privacy, and accountability to prevent biases and misuse of AI technologies. Collaborative Partnerships: Verma stressed the role of global partnerships in advancing AI for development. He called for collaboration between governments, private sector entities, and international organizations to share knowledge, pool resources, and support innovative projects that have the potential to create meaningful impact. Building Local Expertise: Recognizing the importance of local knowledge, he advocated for initiatives to build AI expertise within developing countries. Verma suggested investing in education and training programs to cultivate local talent, which is essential for sustainable AI development and application. AI for Social Good: He provided examples of AI projects that have successfully addressed pressing issues in various regions, such as using AI for early disease detection, optimizing agricultural practices, and improving disaster response. These examples illustrated the potential for AI to contribute positively to society when implemented thoughtfully. Policy and Regulation: Verma called for the development of robust policy frameworks that guide AI deployment while safeguarding against potential risks. He emphasized the need for policies that support innovation while addressing concerns related to data security, ethical considerations, and job displacement. Overall, Alakh Verma's talk was a call to action for a collaborative and inclusive approach to leveraging AI for the betterment of developing countries. His insights aimed to inspire stakeholders to work together to ensure that AI serves as a force for positive change and equitable growth. Alakh Verma

5

🚨🚨 𝗡𝗲𝘄 𝗜𝗻𝗱𝗶𝗰 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗟𝗟𝗠 𝗘𝘃𝗮𝗹𝘂𝗮𝘁𝗶𝗼𝗻 𝗕𝗲𝗻𝗰𝗵𝗺𝗮𝗿𝗸! 🚨🚨 🎉 Excited to share our latest work: 𝗠𝗜𝗟𝗨 - A Multi-task Indic Language Understanding Benchmark, done as a collaboration between AI4Bhārat and IBM Research India, as part of The AI Alliance🌏🤝. 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀: • 85K MCQ questions in 11 Indian Languages. • Spanning 8 diverse domains and more than 40 subjects. • Built with an India-centric approach, evaluating both general and cultural knowledge. We also evaluated 40+ different LLMs (proprietary, open-source, and Indic language-specific). Our key findings: • GPT-4o achieved the highest accuracy at 72%. • Open LLMs (like Llama3.1, Gemma, etc) outperform Indic language finetuned LLMs.  • Models struggle more with culturally relevant domains vs STEM. We hope this benchmark helps in driving the development of more culturally aware and linguistically-competent AI systems for India's 1.4B+ people. 𝗣𝗮𝗽𝗲𝗿 📄: arxiv.org/abs/2411.02538 𝗖𝗼𝗱𝗲 💻: https://lnkd.in/gDhnRsf5 𝗗𝗮𝘁𝗮𝘀𝗲𝘁 🤗: https://lnkd.in/gjruEKjP Work done by: Sshubam Verma Mohammed Safi Ur Rahman Khan Vishwajeet Kumar, PhD Rudra Murthy Jaydeep Sen #AI4Bharat #IBMResearch #NLP #AI #IndianLanguages #Benchmark #LLMs #Evaluation #LLMevaluation #AIAlliance

487

9 Comments

The third article in Volume 8 Issue 1, "The value of novel technologies in context to performance measurement and management: A systematic review and future research directions," is authored by Rajan Kumar V K, Juhani Ukko, Tero Rantala, and Minna Saunila. This paper conducts a systematic review of research on the impact of new technologies on performance measurement and management. It compiles and synthesizes findings from prior studies to offer a comprehensive overview of the state-of-the-art in novel technologies and their application in enhancing performance measurement and management across different industries. Please read this review for free if interested via https://lnkd.in/gK5ZCw2S #systematicreview #literaturereview #researchpaper #phd

14

We are pleased to continue sharing insights from our participation at #ICML2024 in Vienna. A notable research paper by Alexander Wettig, Aatmik Gupta, Saumya Malik, and Danqi Chen has garnered our attention for its exploration of high-quality data selection in language model training. The authors present a novel approach that encapsulates human intuition on data quality by focusing on four key factors: writing style, required expertise, factual accuracy, and educational value. By leveraging language models to perform pairwise comparisons of texts and translating these judgments into scalar values, they propose an efficient method for selecting superior data for model training. Their findings highlight the importance of balancing data quality with diversity, demonstrating that models trained with this approach achieve lower perplexity and improved in-context learning performance compared to traditional methods. This research represents a significant advancement in optimizing language model training, and we extend our gratitude to the authors for their valuable contributions. Read the full paper: https://lnkd.in/dVi3YSgY #ArtificialIntelligence #MachineLearning #LLMs #genAI

21

📕Call for Reading: This #specialissue "Surveys in Algorithm Analysis and Complexity Theory" published 5 high-quality papers, and is viewed by 15955! 👉Thank the Guest Editor Dr. Jesper Jansson's efforts on this Special Issue! 🔗We invite you to read and share: https://lnkd.in/gytzk-tY MDPI Kyoto University #algorithmanalysis #complexity #approximationalgorithm

1

Multimodal foundation models are prone to hallucinations - generating outputs that contradict inputs or lack factual grounding. Our researchers at SCB10X/SCBX Potsawee Manakul and Kunat Pipatanakul in collaboration with University of Cambridge and Tsinghua University, proposed CrossCheckGPT, a reference-free ranking method that correlates highly with reference-based methods. By being reference-free, CrossCheckGPT enables hallucination benchmarking on any target task, and we introduce the first audio-visual hallucination benchmark to the AI community. Check out our paper at https://lnkd.in/gijW6riU to learn more!

43

🌟 Submit your paper now for the #JORS Special Issue on 'Exploring complementarities and differences between Operational Research and Generative AI to Transform Operations' 👉 https://lnkd.in/emdTPCqY 📅 Deadline for submissions is March 2025 #Operations #OperationalResearch #ORMS #AI The OR Society

28

[ 🏆 Best Paper Award] Neha R. (MPRI intern at IRIF in 2023), Eugene Asarin and Mohammed Foughali won the Best Paper Award at FORMATS. This work 𝐟𝐨𝐫𝐦𝐮𝐥𝐚𝐭𝐞 𝐚𝐧𝐝 𝐬𝐨𝐥𝐯𝐞 by an efficient algorithm the 𝐩𝐫𝐨𝐛𝐥𝐞𝐦 𝐨𝐟 𝐪𝐮𝐚𝐧𝐭𝐢𝐭𝐚𝐭𝐢𝐯𝐞 𝐦𝐨𝐧𝐢𝐭𝐨𝐫𝐢𝐧𝐠: compute a real number characterizing to which extent the given trace of a real-time system satisfies its specification. Abstract: Quantitative monitoring mitigates two issues observed in exhaustive, qualitative verification approaches, namely the state-space explosion problem, and the rigidity of their binary verdicts. This is achieved through (i) analysing individual executions instead of building the whole state-space and (ii) providing a robustness measure instead of yes/no answers. In this paper, we consider real-time systems where executions and specifications are modelled as timed signals and Signal Temporal Logic (STL) formulae, respectively. We propose a new temporal robustness measure δ for STL, based on a new distance that we define overtimed signals. In contrast with existing measures, δ provides a precise quantification of distances between the monitored signal and the boundary separating faulty and non-faulty executions w.r.t. an STL property. Thus, δ is suitable for a wide range of real-life perturbations, such as those affecting exclusively a particular time window within a signal. Though we prove that computing δ is NP-hard in general, we provide efficient algorithms for a practical fragment of STL. In particular, this fragment includes the key property of bounded response. 🔗 https://lnkd.in/gsAZuyHV

5

"Modeling with Primary Sources: An Approach to Teach Data Bias for Artificial Intelligence and Machine Learning Education" has received the prestigious Outstanding Design Paper Award at the 18th International Conference of the Learning Sciences (ICLS 2024). The paper was written by COE faculty member Juan Zheng along with J. McClure, F. Bickel, S. Jiang, C.P. Rosé, and J. Chao. This recognition highlights their innovative approach to addressing the critical challenge of data bias in artificial intelligence and machine learning education. The award underscores the quality of their research design and its potential to shape the way data bias is taught and understood in artificial intelligence (AI) and machine learning (ML) curricula. To read the paper, visit https://bit.ly/4cpP1ss.

1

We're excited to share a thought-provoking article by Professor Manish Gangwar, Executive Director of the ISB Institute of Data Science (IIDS) Science, and Dr. Shruti Mantri, Associate Director of the ISB Institute of Data Science (IIDS) The article offers a deep dive into the vulnerabilities in SIM registration processes in India, shedding light on the critical security challenges and potential solutions to safeguard user data... As mobile technology continues to be an integral part of our lives, understanding these vulnerabilities is essential for businesses, policymakers, and individuals alike. Explore their insights on the current landscape and the steps needed to enhance security measures...  Read the full article here and stay informed about the evolving world of data science and security: https://lnkd.in/ggeSwSCz #CyberSecurity #DataScience #ISBInstitute #SIMRegistration #MobileSecurity #IndiaTech

16

In the following article, Professor Manish Gangwar, Executive Director, ISB Institute of Data Science (IIDS), and Dr. Shruti Mantri, Associate Director at IIDS, examine the vulnerabilities in India's SIM registration processes. They address key security challenges and suggest solutions to better protect user data. Read the article to get insights into the current issues and the actions needed to improve security. Click the link to read the full article: https://lnkd.in/ggeSwSCz #iids #isb #ampba #datascience #analytics #ai #cybersecurity Madhu Viswanathan Avik Sarkar, Ph.D. Hemanth Kumar Renu Chaturvedi Peketi Varalakshmi

12

🤖 Why JEPA Proves LLMs are Inefficient for Human-Level Reasoning 🤖 Join Karthikeyan VK from Peninsular Research Operation for a compelling session that explores the limitations of Large Language Models (LLMs) and introduces JEPA (Joint Embedding Predictive Architecture) as a more efficient alternative for achieving human-level reasoning. Learn why LLMs, despite their advanced language skills, struggle with contextual understanding and complex decision-making and see how JEPA offers a framework that more closely mirrors human cognitive processes. Through real-world examples, discover the future potential of JEPA in applications where deep understanding and adaptability are key. Register Now: https://techxconf.com #AI #JEPA #LLMLimitations #HumanCognitiveAI #TechXConf2024 #FutureOfAI #TechXConf

20

Exploring ways to predict product returns in e-commerce by comparing different models, this study co-authored by Dr Biplab Bhattacharjee of Jindal Global Business School (JGBS) found that the Bayesian-optimized XGBoost model performed the best. #research #academics #ecommerce

4

Amazon Web Services announced a substantial investment of more than $500 million in three nuclear power projects across the U.S. #datascience #AI #artificialintelligence https://hubs.li/Q02V-qsn0 If you enjoyed this post, follow me for more insights on AI and data science!

8

The #specialissue "Nature-Inspired Algorithms in Machine Learning (2nd Edition)" has published the fifth paper. We thank the Guest Editors Dr. Szymon Łukasik, Dr. Piotr A. Kowalski, and Dr. Rohit Salgotra for their efforts in this special issue! We invite you to read and share the following: https://lnkd.in/g9zMmabs via MDPI #natureinspiredalgorithms #callforpaper

7