AI & Generative AI Summit @NICE

Artificial Intelligence (AI) has emerged as a powerful tool to enhance customer service and improve customer satisfaction. In various areas of customer service, from self-service and bot interactions to knowledge management and workforce optimization, AI has the potential to revolutionize customer experiences.

This presentation will provide insight into our research activities and showcase the latest trends and developments in the use of AI in customer service operations. Through specific examples, we will discuss how AI can be leveraged to enhance customer experiences and provide superior service.

Join us to learn more about the transformative potential of AI in customer service and discover how we are actively researching and implementing these technologies to provide exceptional customer experiences.

Add to Calendar

Topical Text Classification (TTC) is an ancient, yet timely research area in natural language processing, with many practical applications. The recent dramatic advancements in large LMs raise the question of how well these models can perform in this task in a zero-shot scenario. Here, we share a first comprehensive study, comparing the zero-shot performance of

a variety of LMs over TTC23, a large benchmark collection of 23 publicly available TTC

datasets, covering a wide range of domains and styles. In addition, we leverage this new TTC

benchmark to create LMs that are specialized in TTC, by fine-tuning these LMs over a subset of

the datasets and evaluating their performance over the remaining, held-out datasets. We show that the TTC-specialized LMs obtain the top performance on our benchmark, by a significant margin.

Add to Calendar

Computational clinical psychology is a new and growing field. I will talk about the spectrum between research in NLP versus research in psychology. How can tools from NLP can be used to advance the field of clinical psychology and on the other hand how can tools from clinical psychology be used to advance the field of NLP. To conclude, I will present our last work in which we examined the question - how well do large language models perform on a faux pas tests, a theory-of-mind clinical psychology test.

Add to Calendar

Generative AI tools offer numerous benefits. They expedite R&D efforts, contribute to testing environments, help explore possibilities and support innovation and decision-making. However, using these tools bear risks such as outputs with fake content, false facts and unlawful bias. They may also trigger IP issues and entail confidentiality, privacy and security implications.

The US, EU and other countries are pushing hard toward a formation of regulation around the development and use of AI.

Our talk will focus on the need to establish a corporate framework, to foster a responsible deployment of these tools, harness their benefits, build trust and promote the corporate goals and brand, while mitigating the associated risks.

Add to Calendar

In the talk we’ll introduce our latest development: a novel capability that allows users to interrogate their cloud cost data using natural language. This tool not only enables direct interaction with data but also paves the way for an effortless understanding of anomalies and forecasting results identified by our advanced algorithms. This talk will delve into the mechanics of our breakthrough technology, illustrating its potential in drastically improving cloud cost management and augmenting data accessibility for everyone, regardless of their technical background. Discover how we’re transforming the realm of data analysis, making it more intuitive, efficient, and user-friendly, all with a simple question: “What’s happening with my cloud costs?”

Add to Calendar

Generative AI unlocks a new frontier of opportunities — for knowledge and creative workers— to solve today’s most important challenges. NVIDIA is powering generative AI through a suite of services, pre-trained foundation models, as well as cutting-edge frameworks, optimized inference engines, and APIs to help bring intelligence to your applications. It takes best practices and scale to successfully develop and/or implement Generative AI solutions -  and we will go through some of Nvidia’s offerings on how to best achieve it.

Add to Calendar

Actimize is a company that develops products to help financial organizations detect and prevent money laundering and financial fraud. Due to the increasing sophistication of these financial crimes, there is a need for constant refinement of the technology, with a focus on maximum flexibility and constant learning. To address this, Actimize plans to develop an Online Incremental Learning (OIL) system that tracks data streams and conducts exploratory data analysis to update predictive models based on changes in data and patterns. The innovation in this proposed solution is the truly continuous machine learning that is always online and highly intelligent due to deep statistical analysis, enabling automated building and updating of models in a significantly shorter time. The OIL system will enable a quick response to new types of FinCrime and advanced data mining applications, shortening the analytics construction and improving the effectiveness of risk prediction models.

Add to Calendar

In recent years there were many attempts to compare performance of tree-based ML algorithms and Deep Learning algorithms on tabular data, showing advantage of tree base techniques.

Our long going and extensive research experience at NICE Actimize at the Fraud detection domain is in agreement of the above conclusion – we get optimized performance results while using tree-based ML algorithms. Yet recently we identified an area in which the usage of Recurrent Neural Networks can boost performance of our traditional Fraud Detection by leveraging another level of data which previously was not used in our modeling. In this talk I will present our recent experience leveraging Recurrent Neural Networks in generation of Fraud Detection classification models.

Add to Calendar

Have you ever found yourself in a stressful situation while receiving services from a customer representative? Chances are, you have experienced this just like many others. Join us as we delve into effective strategies for enhancing these experiences, both in real-time and offline interactions. We will explore innovative approaches that empower customer representatives to deliver more positive and efficient services, ultimately leading to improved customer satisfaction.

Add to Calendar

Vector similarity searches are a new type of DB, applicable for many use cases such as anomaly detection, recommendation and search.

In this talk, we would demonstrate building personalizes search engine with Vecsim and CLIP.

Add to Calendar

In the realm of maintenance operations, proactive measures play a crucial role in preventing system failures and optimizing operational efficiency. This talk focuses on the application of deep sequential survival modeling for predictive maintenance. By leveraging advanced techniques such as survival analysis and sequential deep models, we explore how organizations can effectively predict and prevent system malfunctions before they occur.

Add to Calendar

This presentation offers a thorough exploration of deep learning, charting its evolution from the 1960s perceptron model to the cutting-edge GPT model of 2023. By delving into key algorithmic breakthroughs, architectural innovations, and significant models, we provide a comprehensive understanding of the field. This journey through time elucidates how deep learning techniques have profoundly reshaped artificial intelligence and our world.

Add to Calendar

The high-resolution learner activity data that modern online learning environments collect provide a unique opportunity to study ways to improve the pedagogy of interactive learning environments. Applications include adaptive learning algorithms that can provide timely feedback and recommend activities to individual students based on their preferences and state of knowledge; and learning analytics that provide actionable insights about students learning, and can be used to improve instruction, optimize the pedagogic design, and identify students at risk, among other things. My research centers on these directions in MOOCs and K12 learning environments. In the talk, I will present results from studies that use machine learning and NLP to study the behavior of learners, adapt to the needs of individual students, and detect and prevent unauthorized behaviors.

Add to Calendar

The methods for the goodness of fit in classification problems require a prior threshold for determining the confusion matrix. Nonetheless, this fixed threshold removes information that the model’s curves provide, and can be used, for further studies such as risk evaluation and stability analysis. We present a different framework that allows us to perform this study using a parametric PDF.

https://www.oajaiml.com/archive/parametric-pdf-for-goodness-of-fit

Add to Calendar

As the world's aging population grows, Longevity AI steps up to tackle the widening gap between lifespan and healthspan. The current methodologies offer low accuracy due to limited parameters. Longevity AI revolutionizes this approach by harnessing thousands of clinical data points to detect and prevent chronic diseases early. Its groundbreaking AI platform tailors interventions to individual patients' lifestyles, using real-time adaptations and leveraging medical, financial, and behavioral data. Through partnerships with leading healthcare organizations and its multidisciplinary team, Longevity AI redefines the landscape of disease prevention, Envision a prospective reality in which an extended lifespan is combined with well-being.

Add to Calendar

Automatic testing for AI systems is hard, and this challenge becomes even greater when dealing with text and unstructured data. In this talk, we navigate the complexities of testing text-based AI models. We will explore automated testing methodologies for such models & data, discuss how to combine this with domain knowledge, spotlighting the challenges, solutions, and best practices using the Deepchecks NLP Open Source testing package. The talk is suitable for various levels of participants, from hands on practitioners with deep knowledge about AI & NLP to managers with only general knowledge about the space.

Add to Calendar