Data Science and Governance – Pagina 6

The (Un)reliability of Saliency methods – Google Research

Massimo2021-07-160713 views

In the exploration of deep model interpretation, saliency methods emerge as a popular technique for evaluating feature importance. They assign importance scores to input features, indicating their utility in model performance. High scores suggest significant performance degradation in their absence. However, investigations, such as those by Google Research, reveal the inherent unreliability of these methods. The crux of the issue lies in their sensitivity to non-influential factors and failure to maintain input invariance, leading to potentially misleading attributions. This challenges the effectiveness of saliency methods in providing accurate explanations of deep learning behaviors.

Speaker at “AI, Data Analytics & Insights Summit – DACH”, 11th – 12th November 2021

Massimo2021-07-0601,1K views

At the upcoming “AI, Data Analytics & Insights Summit – DACH” on 11th – 12th November 2021, a session will be dedicated to exploring Artificial Intelligence applications within Research and Development. This interactive, senior-level online meeting will convene 250 experts from the DACH region, offering a unique platform for sharing insights and advancements in the field.

What I learnt talking to people and speaking at Conferences

Massimo2021-07-020591 views

In effective communication, the essence lies in simplicity and relevance. Overloading information overwhelms, whereas tailoring content to the audience’s interests and preferences enhances retention and engagement. This approach, termed listener-centered communication, pivots away from speaker-centric narratives, focusing instead on what resonates with the audience. By initiating conversations that address current challenges and priorities, one can craft messages that are both compelling and concise. Leading with benefits rather than personal burdens ensures the audience’s attention is captured, paving the way for productive discourse. This methodology advocates for a strategic, audience-aligned communication, emphasizing the power of delivering precisely what is necessary, no more.

Leveraging NLP in Knowledge Management: a Case Study of Lab Document Management

Massimo2021-07-010546 views

In a pioneering effort to streamline laboratory knowledge management, a sophisticated system leveraging Natural Language Processing (NLP) and machine learning models, including BERT and GPT, was developed to efficiently manage a massive repository of scanned documents. By applying advanced techniques such as topic modeling, document clustering, and semantic similarity analysis, this system significantly improved document accessibility, categorization, and retrieval. The creation of a detailed ontology, integrated with public data sources, further enhanced data interoperability and research collaboration, showcasing the transformative potential of NLP in handling complex data landscapes.

How many fields in data science?

Massimo2021-06-2701,K views

Data Science is a research activity… mostly Data-driven scientific discovery is regarded as the fourth science paradigm The twenty-first century has ushered in a new age that is coined as data science and big data analytics. Data-driven scientific discovery is regarded as the fourth science paradigm. Data science has been a core driver of the new-generation science, technologies and economy, and is driving new researches, innovation, profession, applications and education across both disciplines and business domains. There are many scientific and technical challenges associated with big data, ranging from data capture, creation, storage, search, sharing, modeling, representation, analysis, learning, visualization, explanation, and decision making. Among the many data characteristics and complexities to be addressed, I mention the hybridization of heterogeneous, multisource, hierarchical, interactive, dynamic,…

Speaker at FUTURE Labs 2021

Massimo2021-05-250742 views

At FUTURE Labs 2021, the spotlight on Artificial Intelligence’s role in Research and Development underscores its pivotal contribution to shaping the laboratories of tomorrow. The conference, renowned for its diverse assembly from academia to industry giants across various sectors, including Biotech, Pharma, and more, serves as a crucible for innovation. It invites a confluence of ideas and visions, aiming to redefine laboratory operations and efficiency. With discussions spanning nine crucial themes, including AI & Machine Learning, Digital Transformation, and Data Management, the event promises a comprehensive exploration of the technological forefront, all delivered in English, facilitating a global discourse.

Microservices architecture: the case of AWS

Massimo2021-04-120680 views

Serverless computing represents the pinnacle of cloud abstraction, focusing solely on code functionality rather than underlying infrastructure. It eliminates concerns about servers, operating systems, and runtime environments, allowing for the execution of code snippets upon specific events. AWS Lambda exemplifies this model, offering a platform where only the necessary code runs when triggered, devoid of server or OS knowledge. This approach is instrumental in developing microservices architectures, with AWS services like API Gateway and S3 acting as event sources to invoke Lambda functions. Such architecture simplifies operations, reducing maintenance and enabling a focus on application logic, thereby enhancing efficiency and scalability in deploying Artificial Intelligence solutions.

Speaker at Swiss Data Leaders Conference

Massimo2021-03-310705 views

At the SWISS LEADERS CONFERENCE, an exchange of innovative ideas on data strategy and digital transformation took place. The focus was on leveraging Big Data Analytics and AI to foster digitalization, enabling service-oriented business models, and driving market innovation. Investments in skilled teams, software, and architecture have led to exciting use cases, with data-driven startups notably disrupting industries.

Speaker at the next Future Labs 2021

Massimo2021-01-2901,1K views

In the realm of innovation, Future Labs Live 2021 stands as a beacon, uniting over eighty global experts in a comprehensive dialogue on the forefront of digital transformation, data science, Artificial Intelligence, and Machine Learning. This event, spanning two days, is pivotal, addressing the urgent need for technological, organisational, and cultural shifts across industries.

Interpreting deep learning models

Massimo2020-07-1701,K views

Interpreting deep learning models is crucial for diverse applications such as healthcare and self-driving cars. Understandably, errors can have catastrophic consequences. Thus, achieving interpretability is essential for decision-makers. Properties like fidelity, comprehensibility, and accuracy are vital for evaluating interpretability. Various methods, including visualization techniques and knowledge distillation, offer insights into complex models. However, quantifying interpretability remains a challenge. For more information, refer to research papers on refining deep neural networks and interpreting CNNs. Enhancing interpretability not only fosters trust in AI but also mitigates risks in decision-making processes.

Why Meta-learning is important

Massimo2020-03-090758 views

Meta-learning, a groundbreaking approach in AI, empowers machines to rapidly adapt and learn from minimal data. By transcending traditional machine learning, meta-learning revolutionizes various sectors like healthcare, finance, and education. This technique facilitates few-shot learning, enabling models to excel with limited examples, a paradigm shift from data-intensive methods. Meta-learning’s impact spans diverse domains, from personalized education to drug discovery in pharmaceuticals, promising accelerated innovation and optimized processes. Embracing meta-learning heralds a future where AI systems dynamically evolve and excel in novel tasks with unprecedented efficiency.

AGROVOC: An Introduction to the World’s Largest Agricultural Vocabulary

Massimo2018-07-2001,3K views

Discover AGROVOC, the concept scheme revolutionizing agriculture. Learn how it organizes agricultural domains, differentiating concepts and terms while harnessing the power of RDF. Uncover its potential through the AGROVOC RDF file, enabling seamless integration into diverse applications and systems.