蘑菇视频

Melanie Laffin discovered her passion for technology in high school. Despite knowing the challenges women face in the field鈥攁nd the efforts of some to dissuade her from making it her career鈥擬elanie was undeterred and chose to chase her dream.

Today, Melanie works in the burgeoning fields of artificial intelligence (AI) and natural language processing (NLP). She is using her knowledge to support and raise the profile of other women, including in a groundbreaking project for the Smithsonian鈥檚 Office of the CIO as part of supporting efforts for the Smithsonian American Women鈥檚 History Museum (SAWHM).

SAWHM is one of the Smithsonian Institution鈥檚 many museums. 鈥淚鈥檓 a huge Smithsonian fan girl. It鈥檚 one of the reasons I moved to the Washington, DC, area,鈥� Melanie says. So, when 蘑菇视频 Allen was compiling a team to support a Smithsonian project to identify and recognize unknown contributions by women, she immediately raised her hand.

Here Melanie shares highlights from the project and her work at 蘑菇视频 Allen, along with her thoughts on harnessing AI as a force for good.聽

In areas from botany and entomology to art and antiquities, women have played a significantly larger role than they鈥檙e currently given credit for in the Smithsonian鈥檚 vast collections.

For example, in the 19th and 20th centuries, a female donor鈥檚 name might be listed under her husband鈥檚 name or family estate. And many women who participated in research projects or scientific expeditions were often excluded from official documentation.

The truth about these 鈥渉idden鈥� contributors lies in the millions of records the institution has collected over the years鈥攔ecords far too vast to sift through and cross-reference by hand. The Smithsonian Digitization Program Office is now digitizing as many of these records and objects as possible. It鈥檚 a multiyear undertaking that involves collaborating with the institution鈥檚 individual museums.

This is where Melanie and her 蘑菇视频 Allen team of experts come in. Their goal was to use AI tools like NLP to extract the names of individual women from the Smithsonian鈥檚 collections and connect them to their contributions, fields of expertise, and other individuals.

Melanie and her team started their 90-day pro bono project鈥攃alled Hidden Impacts鈥攚ith a trove of collections metadata and digitized Smithsonian annual reports from the 1880s to the 1960s. Using NLP techniques like topic modeling and named entity recognition, they were able to extract names, donation types, dates, and other affiliations.

The individual names they extracted via predictive techniques often spurred fresh questions. For example, were 鈥淎. Smith鈥� and 鈥淎lex Smith鈥� the same person, and what gender did this person identify as? To narrow in on a probable answer, the team conducted analyses such as gender estimations, mapping contributor names against data like Social Security records and popular names for a specific year.

At the end of the engagement, she and her team had built a searchable network graph showing relationships between women, other collaborators, institutions, and collections鈥攁nd unearthing previously hidden networks and collaborations.

It鈥檚 a powerful first step for further discovery. 鈥淲e鈥檙e building the scaffolding for future research,鈥� Melanie says. 鈥淔rom this work, researchers and subject matter experts can fill in gaps to identify uncredited contributors to the Smithsonian collections.鈥�

The Smithsonian is just one of many organizations that has benefitted from Melanie鈥檚 passion and expertise during her time at 蘑菇视频 Allen.

鈥淚 want a world where robots do all the boring, repetitive stuff for us. That way, we can spend our time doing the not boring, repetitive stuff,鈥� she says.

To help 蘑菇视频 Allen鈥檚 government clients work toward this future vision of AI, Melanie explores new applications like cognitive automation, which combines AI with process automation to improve business outcomes, and fresh approaches to pressing challenges. One example of a new approach: 鈥淒uring the COVID-19 pandemic, some agencies received a flood of applications filled out with complete gibberish, causing huge backlogs,鈥� says Melanie. 鈥淲hat about an NLP detector to remove those from the queue?鈥�

Right now, Melanie is working with a major federal benefits provider on fraud prevention services, using AI to both stop criminals and improve the user experience.聽

鈥淚f my career has a theme, it鈥檚 to be as helpful to people as possible,鈥� Melanie says. 鈥淗aving a toolset to fix problems and create something鈥攆or me, that鈥檚 so fulfilling. 蘑菇视频 Allen has given me the opportunity to turn complex, messy systems into things of beauty.鈥澛�

Melanie says she and her team鈥檚 work on the Smithsonian鈥檚 Hidden Impacts project is just beginning. She plans to submit the research to data science conferences and hackathons, especially those focused on engaging women in technology careers, as well as 蘑菇视频 Allen鈥檚 own Women in Data Science incubator series. 鈥淲e want as many people as possible working on it,鈥� she says.

For Melanie, projects like these are all part of a larger vision: helping people see AI鈥檚 potential as a force for good.

She emphasizes that the necessary caution over AI鈥檚 risks shouldn鈥檛 overshadow the many benefits of this technology: to help people save time, make the most of data, grasp connections, and more. And the Smithsonian鈥檚 Hidden Impacts project is a prime example.

鈥淎I is good at finding hidden patterns and helping us see the actual truth,鈥� Melanie says. 鈥淎I uncovers what鈥檚 really going on, like when credit hasn鈥檛 always been given to the right people. We want to help give credit where credit is due.鈥澛�