Steph Buongiorno

Text Mining for Historical Analysis

Wed, 11 Oct 2023 12:01:35 +0000

Democracy Viewer

Tue, 10 Oct 2023 12:01:35 +0000

Mapping the Elusive: Using Network Analysis to Understand Slavery, Debt Relations, and the Emergence of a Free Population of African Descent in 18th-Century Colonial Havana

Mon, 09 Oct 2023 00:00:00 +0000

Who is the Working Class, Socialist James Bond? Exploring Propaganda through Knowledge Graphs and AI-Assisted Forensic Reading in Bulgarian Spy Fiction

Mon, 11 Sep 2023 00:00:00 +0000

Interactive Gallery

Click "Load Graph" to render a visualization.

The Momchilovo Affair Network Graphs

The Momchilovo Affair Knowledge Graph 1

Graph not loaded

The Momchilovo Affair Knowledge Graph 2

Graph not loaded

The Momchilovo Affair Knowledge Graph 3

Graph not loaded

The Momchilovo Affair Knowledge Graph 4

Graph not loaded

The Momchilovo Affair Knowledge Graph 5

Graph not loaded

The Momchilovo Affair Knowledge Graph 6

Graph not loaded

The Momchilovo Affair Knowledge Graph 7

Graph not loaded

The Momchilovo Affair Knowledge Graph 8

Graph not loaded

The Momchilovo Affair Knowledge Graph 9

Graph not loaded

Topic Model - The Momchilovo Affair and Mr. Nobody Novels

Graph not loaded

Mr. Nobody Network Graphs

Mr. Nobody Knowledge Graph 1

Graph not loaded

Mr. Nobody Knowledge Graph Network 2

Graph not loaded

Mr. Nobody Knowledge Graph Network 3

Graph not loaded

Mr. Nobody Knowledge Graph Network 4

Graph not loaded

Mr. Nobody Knowledge Graph Network 5

Graph not loaded

Mr. Nobody Knowledge Graph Network 6

Graph not loaded

Mr. Nobody Knowledge Graph Network 7

Graph not loaded

Mr. Nobody Knowledge Graph Network 8

Graph not loaded

Mr. Nobody Knowledge Graph Network 9

Graph not loaded

Mr. Nobody Knowledge Graph Network 10

Graph not loaded

Mr. Nobody Knowledge Graph Network 11

Graph not loaded

Mr. Nobody Knowledge Graph Network 12

Graph not loaded

Mr. Nobody Knowledge Graph Network 13

Graph not loaded

Topic Model - The Momchilovo Affair and Mr. Nobody Novels

Graph not loaded

Topic Models

Topic Model - The Momchilovo Affair and Mr. Nobody Novels

Graph not loaded

Topic Model - Whole Corpus of Twenty Nine Novels

Graph not loaded

Foundations and Applications of Humanities Analytics

Sun, 10 Sep 2023 12:01:35 +0000

As the teaching fellow for Foundations and Applications of Humanities Analytics (2022-2023) at the Santa Fe Institute, I design and teach workshops for participant researchers. I foster an inclusive and intellectually rigorous classroom for researchers with diverse backgrounds, from senior faculty to early career graduate students.

You can view our repository on GitHub for data sets and code examples that we have used in the past.

These courses are supported by National Endowment of the Humanities (NEH) Grant (no. HT-272418-20)

Database Escrituras Protocolos 1640 a 44 y 50 and 1730 a 1733

Sat, 09 Sep 2023 12:01:35 +0000

Transcripts of notarial records from the Fondo Escribanías in the National Archive of Cuba, an endangered repository. The transcripts document pawnship practices (or a credit instrument known in the historical record as retroventa) involving enslaved people in Havana, Cuba during the 17th and 18th centuries. These records are important for analyzing how the enslaved person functioned as a “social connector,” linking a wide range of creditors and debtors through contracts in the colonial urban economy. The pawning of enslaved individuals, whose value was sometimes divided in multiple parts and to multiple creditors, served to hedge risk and provided temporary access to enslaved labor, at a time of labor scarcity, when the human trade to Cuba declined and several epidemics devastated the island. Over time, as ownership became increasingly fragmented, some enslaved people started buying stakes in themselves.

https://doi.org/10.7910/DVN/AZCY4B

https://doi.org/10.7910/DVN/AEDRPT

Word Embeddings as a Key to the Study of Bias, Race, and Gender in Congress, 1880-2010

Sat, 09 Sep 2023 12:01:35 +0000

The Congress Viewer Demo App

Thu, 07 Sep 2023 12:01:35 +0000

This prototype web app, the Congress Viewer , explores how the tools to measure lexical change – and more sophisticated NLP strategies such as parsing and measuring grammatical relationships – can open a window into political discourse that will make what happens in Congress more transparent to researchers and everyday citizens alike.

The Congress Viewer is a prototype app. For our final app, citizens using our toolset will be able to easily navigate from an overview showing change over time, to a dramatization of how different candidates talk about the same issue, to the in-text mentions of word use that a computer has used to produce the visualizations in question.

Below is a truncated version of the app. Click here to interact with a full sized version.

The Congress Viewer, a prototype text mining app, demonstrates how the tools used to measure lexical change, as well as sophisticated NLP strategies such as parsing and measuring grammatical relationships, can open a window into political discourse. On the one hand, it can make Congress more transparent. On the other, these tools can be used to gain new insights into the evolution and nature of political language as it occurs in different contexts, such as different time periods and discourse communities.

Beyond the Black Box: Toward Transparent AI for Computational Text Analysis in the Digital Humanities

Thu, 10 Aug 2023 00:00:00 +0000

PANGeA

Tue, 08 Aug 2023 12:01:35 +0000

Large language models (LLMs) offer unprecedented flexibility in procedural generation, enabling the creation of dynamic video game storylines that evolve with user input. A critical aspect of realizing this potential is allowing players and developers to provide dynamic or free-form text to drive generation. Ingesting free-form text for a video game poses challenges, however, as it can prompt the LLM to generate content beyond the intended narrative scope. In response to this challenge, this research introduces Procedural Artificial Narrative using Generative AI (PANGeA) for leveraging large language models (LLMs) to create narrative content for turn-based, role-playing games (RPGs).

PANGeA is an approach comprised of components including a memory system, validation system, a Unity game engine plug-in, and a server with a RESTful interface that enables connecting PANGeA components with any game engine as well as accessing local and private LLMs. PANGeA procedurally generates level data like setting, key items, non-playable characters (NPCs)), and dialogue based on a set of configuration and design rules provided by the game designer. This process is supported by a novel validation system for handling free-form text input during game development and gameplay, which aligns LLM generation with the narrative. It does this by evoking the LLM’s intellegence to dynamically evaluate the text input against game rules that reinforce the designer’s intent. To enrich player-NPC interactions, PANGeA uses the Big Five Personality model to shape NPC responses.

To explore its broad application, PANGeA is evaluated across two studies. First, this research presents a narrative test scenario of the prototype game, Dark Shadows, which was developed using PANGeA within the Unity game engine. This is followed by an ablation study that tests PANGeA’s performance across 10 different role-playing game scenarios–from western to science fiction–and across three model sizes: Llama-3 (8B), GPT-3.5, and GPT-4. These evaluations demonstrate that PANGeA’s NPCs can hold dynamic, narrative-consistent conversations that, without the memory system, would exceed the LLM’s context length. In addition, the results demonstrate PANGeA’s validation system not only aligns LLM responses with the game narrative but also improves the performance of Llama-3 (8B), enabling it to perform comparably to large-scale foundational models like GPT-4. With the validation system, Llama-3 (8B)’s performance improved from 28% accuracy to 98%, and GPT-4’s from 71% to 99%. These findings indicate PANGeA can help game designers generate narrative-consistent content while leveraging LLMs of different sizes, suitable for various devices.

Steph Buongiorno [Corresponding Author], Jake Klinkert, Tanishq Chawla, Zixin Zhuang, and Corey Clark. “PANGeA: Procedural Artificial Narrative Using Generative AI for Turn-Based Video Games.” Proceedings of the 2024 AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE), Lexington, KY, USA, 2024.