This reusable LSTM-based household diagnostic decreases misdiagnoses of CVDs (the cause of 32% of deaths) and can reduce thousands of tons of medical waste daily. The 83.8% diagnosing accuracy and 93.3% overall functionality show promise in detecting various diseases and creating a cleaner world.

The Wolfram System's recent advances in disciplined convex programming are applied to a multi-objective optimization problem: find a cubic spline function with small data distance and curvature. We add positivity, monotonicity and convexity constraints, which we express as norm cone constraints.

This talk will introduce new upcoming chemistry functionality for Wolfram Language, including enhancements for molecular visualization and new data structures to represent crystal structures and large biomolecules. In addition, the ChemistryFunctions paclet will be presented, with functions for dynamically manipulating molecular structure, enhanced substructure matching and visualization, as well as dynamic controls and predicate functions.

Wolfram Language provides an advanced framework for rapidly prototyping neural network applications and exploring research ideas. In this talk, I demonstrate the language features that helped me develop a novel neural network that learns compact non-differentiable Boolean functions.

In this presentation we discuss how both theoretical and software engineering constructs made for epidemiological compartmental models (ECMs) can be "upgraded" into war system dynamics (SD) models. Several types of analogies are drawn, both operational and economic. The most interesting is the turning of a geo-spatial-temporal ECM into a war SD model. The presentation uses numerous "big picture" workflow diagrams and breakdowns. Concrete models are demonstrated and explained.

The complex imaging workflow associated with astrophotography can be automated using Wolfram Language's intrinsic image processing and statistical data manipulation functionality. Astronomical imaging involves taking a sequence of exposures that have to be calibrated, aligned, stacked, denoised and then stretched to obtain any sort of image. Color imaging involves following this pipeline for each of three or four color channels that then are combined to form the final image. This talk illustrates this workflow with galaxy images taken in May of 2023 from a light-polluted location in Urbana, Illinois. The image of M101 also shows a supernova SN 2023ixf that was discovered the week prior.

I will describe our experience at the University of California, Irvine, teaching mathematical methods to physical science students at the undergraduate and graduate level. The classes utilize an ebook text that is a collection of Mathematica notebooks. The classes are paperless: the assignments are written in Mathematica notebooks, and the students do their work and submit a notebook. For the undergraduate version, we are experimenting with a web interface that automatically grades reasonably complicated symbolic input using calls to Wolfram Cloud APIs.

Learn about book publishing using Wolfram Language. We'll explain our various programs for supporting authors working to publish both with Wolfram Media and with other publishers. A brief demonstration of the WolframBookTools palette for authoring will be given. We will leave time for Q&A.

Mancala is a two-player board game known from antiquity and has many variations of play worldwide. We studied the most common North American version known as Kalah. The traditional board consists of six pits, each containing four seeds, and an empty store for each player, and the objective is to capture the greatest number of seeds (one's own and the opponent's) in one's store by moving them from one of the pits and sowing them in succeeding pits and the player's store. The game involves both counting and strategy, as the rules allow extra turns. We made use of game graphs and the tools of multiway analysis to understand how the complexity and "fairness" of the game changes as the board configuration and choice of rules are varied. As the score determines the winner, we also examined how early in the game a win can be forced. Two of the tools we used to study mancala were a function to display the game board for any given state of play and another function to sow the seeds from a selected pit and follow the rules for capture and extra turns. It was then a simple matter of turning the displayed board into a dynamic interface for two players that permitted a fully interactive game to be developed.

Over the past years, we have developed several tools for actually building code representing objects that are often easier to think about without looking at syntax—for example, GeometricScenes and Entities, which both are powerful tools in Wolfram Language, but require involved syntax in order to be specific enough to work as functions. In this talk, we will give an overview of these new functionalities and discuss the ways these tools will make advanced Wolfram Language functionality accessible to newcomers and a younger audience, and are prime examples of the flexibility of the notebook interface.

Wolfram AI platform is used in medical imaging at Deakin for its ease of use with non-programmer students and staff. Successful research projects include DNN detection and staging of COVID-19 from chest x-rays, and the detection and localisation of prostate cancer in magnetic resonance diffusion images.

Misinformation and other attacks on the Sustainable Development Goals and ESG investing have increased dramatically in the past year. We use new Wolfram LLM tech as well as other core functionality—entities, WebSearch, Wikipedia, WebSessions—to build tools that help address these issues.

Robotics, as an area of teaching and research, can be fascinating and intimidating at the same time. The kinematics of robots involve formulating and solving systems of algebraic or trigonometric equations. Their dynamics and control require derivation and analysis of coupled nonlinear ordinary differential equations. These computations can be so demanding that they may impact the students' motivations to learn. They also pose great challenges to researchers in their efforts to make fresh contributions to this field. In this context, a robust and powerful computer algebra system such as Wolfram Mathematica can be the best friend of a teacher or a researcher. The speaker has spent 16 years in the first role and nearly 25 years in the second. During this time, he has developed several symbolic codes to solve benchmark problems in the field of parallel manipulators. He has also created computational workflows and demonstrations which aid the students at Indian Institute of Technology Madras to analyse and understand complex behaviours of robotic systems with much greater ease and confidence. In this talk, the speaker will share a few of these developments.

Mathematica is a powerful computing environment that was initially used in mathematics and physics, but is now gaining popularity in various disciplines. It offers numerous built-in functions for financial computations, simulation, optimization, risk management and forecasting. Despite its capabilities, it is not widely used in financial data analytics. The presentation demonstrates the use of important built-in financial functions and custom functions for financial data analytics using real-world financial data. Examples cover various finance topics, and all results are reproducible without modifications.

This talk will discuss how the advanced neural network framework within Wolfram Language was used to train and develop a model built from simulation-based data that went on to be used for real-time monitoring of hydrate formation risk at an LNG (liquified natural gas) facility.

This presentation demystifies the Query function for the intermediate user and shows how users at all levels can effectively use it as a functional programming construct. It draws from parts of the author's new Wolfram Media book, Query: Getting Information from Data with the Wolfram Language.

We present updates to the automated geometric functionality of Wolfram Language, including such existing system functions as GeometricScene, RandomInstance and FindGeometricConjectures, as well as such resource/future functions as FindGeometricProof and GeometricSolveValues.

Mathematica and Wolfram Language can connect to basically every open or commercial database out there—relational (SQL) or NoSQL flavors too. From small to big companies, they have relational database management systems to support CRM, ERP and other enterprise systems. Lots of that data is just sitting there waiting to be explored with a multiparadigm data science approach. Interesting patterns can be found in business data and can be combined with Wolfram-curated data and internet-available datasets too. We will show how to build visual GUI tools to connect to relational databases and navigate visually on database objects.

Rational zeta series are series in which the n summand is a rational multiplied by a zeta value. We give close formulas to compute a family of these series. Currently no software can reproduce our results.

Most Wolfram Language programs consist of a set of functions, each of which transforms its inputs into outputs, which are then "plugged into" each other to create a nested directed acyclic graph. I will present an alternate front end that lets users work with this graph directly in a zoomable map-like interface.

This talk introduces the significant enhancements in the upcoming redesign of the ServiceConnections framework. Discover a streamlined process for creating new API connections, with features such as automatic request generation and response parsing, automatic parameter completion and verification, and versatile authentication methodologies. Join us for a technical deep dive into the next wave of API integration.

StyleGAN (Style Generative Adversarial Network) is a type of generative model capable of generating realistic images. It is not trivial to explain what happens inside. In this computational essay, I propose an explanation based on active shape and appearance models.

Many terrific computational stories start out as notebooks, but less commonly do they evolve into compelling, ongoing narratives. Why is this? The challenge of extending computational essays into robust computational systems often presents as a step too far. But is it?

Symbolic integration, a cornerstone of computational mathematics, traditionally relies on intricate algorithms to find antiderivatives. Rubi, a rule-based integrator, takes a different and unique approach. Instead of algorithms, Rubi uses thousands of human-readable mathematical rules, systematically applied, to determine integrals. This method provides more transparent integration steps and often yields simpler and more accurate results than traditional means. In this talk, we'll delve into the world of Rubi, exploring its potential and comparing it to conventional integrators.

This talk will cover the use of LLMs and AI in enhancing Wolfram's built-in food data. We will show examples of how to use Chat Notebooks and the Wolfram ChatGPT plugin to retrieve nutrition data and analyze it. We will demonstrate how LLMs can transform recipe text into structured datasets in Wolfram Language from which we can generate graphs. We will also demonstrate new Wolfram Function Repository functions for food that can be easily used by chatbots to complete tasks for the user.

NotebookTemplating is a paclet for creating repeatable documents using templates. This talk will give a brief overview of how the paclet works, focusing more on how Wolfram and other companies have used report generation for better workflows.

Reconstructions of the geological history of the Earth can be carried out using a number of tools in Wolfram Language. GeoGraphics, EntityValue and the Wolfram Function Repository provide a variety of tools that aid in exploring historical geology and will all be investigated in this presentation.

Imagine Galileo and Descartes collaborated on the art of falling. Galileo proffered time as range, making space as domain arbitrary. Domain fall rate not of fixed meter presented a challenge. They used algebraic tiles to discover uniform acceleration spacetime. I use their first-second tile to construct our first central force spacetime square inquiry.

New taxonomic species data collections encompass ecological, geographical as well as biological knowledge of living organisms including animals, plants, fungi and other microorganisms. With the new hierarchical entity type syntax, one can access general taxonomic information of species as well as categorized characteristics of organisms specific to each kingdom of life. Improved data lookup support allows in-depth exploration of diversity of life. 

Discover UpLift, a next-generation gas lift simulator crafted in System Modeler V13.3 and enhanced by Mathematica. Beyond petroleum engineering, UpLift offers a deep dive into real-world oil production scenarios using authentic well data. Experience the future of simulation, bridging technology and nature.

In a recent Community post, tools for calculating a hundred thousand tetrahedron centers were used to make an animation. For this talk, a dozen new geometry tools for both 2D and 3D spaces will be discussed.

Protein visualization plays a pivotal role in understanding the intricate world of biomolecular structures. In this talk, I will discuss protein visualization in Wolfram Language, with a specific focus on the chemistry-related functionalities and applications of the Wolfram ProteinVisualization paclet. Additionally, I will share examples illustrating how this paclet can be effectively utilized in both education and research.

Visual schematic representations have built concrete intuition and aid in intuitive comprehension. Despite the enormous progress that modern abstract geometry has made since the late nineteenth century, the property of concentric circles—where they touch each other at the imaginary circular points at infinity—is still described as STRANGE in geometry books published in the twenty-first century. Some mathematicians may argue, "Since equations prove a theorem is correct, we do not mind if the visuals are incomplete or unavailable in the first place." However, a tradition dating back to the Greek era dictates that we describe the existence of a mathematical object geometrically and only consider it a mathematical reality when a purely geometric interpretation becomes possible. Even today, intuitive ways to grasp various facts and problems in mathematics are becoming increasingly vital. Presenting imaginary and infinitely distant elements using Mathematica's advanced pattern matching and rewriting capabilities should contribute to a correct understanding of complex geometry for many people, not just specialists.

Exploring the SIR model's utility in understanding infectious diseases, this study delves into its fundamentals and application. Using the 1960s Hong Kong flu outbreak in NYC as a case study, it highlights the model's role in disease dynamics and intervention measures.

We are going to study finite lattice systems in the context of simple rewriting rules: based on discrete manifolds and Cayley graphs, in this project we look for the set and taxonomy of simple rewriting rules that preserve certain properties that can later be understood as conserved measurable magnitudes of a physical system, like the Brouwer degree or winding number. By investigating those toy models about field configurations of discrete fiber bundles, we will later apply local perturbations to a field configuration, looking for discrete analogs to Noether's theorem. Furthermore, via coloured graphs, with this approach we will aim to characterise (energetically stable) physical-inspired systems in the context of many-particle quantum finite distributions in highly symmetric arrays or low-dimensional spin-like states as domain wall in magnetization theory, being local solutions of minimal information configurations. Even mesoscopic scale is responsible for exotic topological defects; with this discretization, we will try to understand the computational first-principle mechanism that allows nontrivial winding numbers and the material and geometries to host pseudo-particle structures like skyrmions, bubbles or merons. We think that this inductive reasoning here would offer a valuable approach for finding new phenomena or even reinterpreting old ones.

This year I will continue my problem-solving style shown in '22 and demonstrate the problem-solving skills in these four domains: geometry, combinatorics, number theory and algebra/analysis. I have selected couple of problems and examples from the following published resources:
R. Honsberger's From Erdös to Kiev: Problems of Olympia Caliber (or related Crux volumes)
M. H. Weissman's An Illustrated Theory of Numbers
S. Savchev and T. Andreescu's Mathematics Miniatures
R. Stanley's Enumerative Combinatorics

Wolfram Language's built-in functions and Wolfram Function Repository resources have a unique charm to make the exploration of the solutions very expressive. It is critical to see the connection between the "Aha!" moment and solid examples generated by Wolfram Language. It is my wish for all my audience to enjoy the talk and start their own journeys of solving challenging problems with Wolfram Language.

This presentation outlines how the Wolfram technology stack can be utilized to create workflows that maximise the potential of emerging large language models such as OpenAI's GPT, Google's Palm, Meta's LLaMA and more. Special attention will be paid to processing and analyzing large unstructured data such as those present in Sustainable Development Goal reports or conference video collections. This demonstration builds on previous presentations by the authors to highlight new technologies while emphasizing familiar methodology. Specific topics covered include:
*Overview of tools and options available to Wolfram users
*The pros and cons of various methods for accessing LLM outputs
*Methods for ingesting and processing large quantities of unstructured data for use in LLMs
*Transcripts from videos
*Working with PDFs and other unstructured data sources
*Interfaces between LLM models and Wolfram outputs
*All-in-one tools to create content and summaries that maximize the strengths of internal and external resources
*Tips and tricks to improve results