MARVL

Over five million U.S. public school students don’t have adequate mastery of English. In addition, they may not be fluent or literate in their home-language. School districts recognize this problem and are working to solve it. They train teachers and provide ESL coordinators who come in once or twice a week. But this isn’t enough to help get ELLs to where they need to be by semester’s end. MARVL is a fun learning platform, easy to use, available on phones or tablets, that kids can use alone or together with their teachers and parents. With MARVL, students learn English words simultaneously with their home language. We are currently focused on Spanish, Haitian Creole, and Korean, with plans to add more languages, especially dialects and at-risk languages. Scaling means aiding global bilingual education, resulting in more students succeeding in school and having a widened selection of career paths.

One in four students in the United States learns English as a new language when they begin school. These students, ELLs, face pervasive achievement inequities. This is attributed in part to the challenge of acquiring a new language while also learning content. ELLs must rapidly acquire thousands of English words to comprehend grade-level texts (August & Shananhan, 2006) and schools fail to provide optimal instruction that fully addresses their needs. An increasingly diverse student population (with over 300 home languages spoken), ELL students experience achievement inequities, especially when it comes to reading comprehension and STEM education. COVID has only further intensified these achievement inequities.    

Decades of research indicates that continued support in a child’s first language, in addition to English, results in higher achievement outcomes (Burchinal et al., 2012). However, the U.S. has a bilingual and ESL teacher shortage and schools struggle to provide optimal instruction to help our ELLs access curriculum in their home language while also learning English. In addition, most of the existing software for supporting ELLs are computer-based and therefore only can be used in schools, which rarely have enough computers for all students. And few low-income families have laptops or desktops.

MARVL enables K-12 ELLs to grasp concepts taught in the classroom quickly and easily, especially within STEM subjects. It provides interactive, individualized instruction, in both the user’s home language and English, in an immersive environment. It leverages AR and AI to provide differentiated, immersive learning experiences for students. By scanning vocabulary cards, users trigger a learning experience helmed by 3D characters who fully engage with them, delivering instructions, encouragement, and demonstrating concepts. We strive to parallel a learning experience obtained from in-person bilingual teachers who learners hold conversations with, and mimic the process of acquiring a first language.

Our solution goes beyond traditional instruction. We leverage both audio (spoken definitions, expanded descriptions, sounds) and visual (multiple pictures, 3D animations, videos) capabilities to enhance learning. Animations, sound effects, and texts emphasize letter-sound relationships, and our animated characters prompt vocal participation, holding students’ attention and retention, adding excitement and gamification. Users can use the solution at school or home, removing issues that teachers face when dealing with a limited class time and large class sizes. It works on any smartphone or tablet, is easily customized, can be aligned with existing curricula, and most of all, can be offered in many different languages.

The fastest-growing segment of the U.S. K–12 public school population, ELL students are estimated at 5.5 million, representing 11% percent of all K–12 students. ELLs are a highly diverse student population. The majority of kindergarten-aged ELLs, about 85%, are born in the U.S. (Park & Batalova, 2015), however, the U.S. also has older ELLs who arrive with limited or interrupted education backgrounds and varying first language literacy backgrounds. Due to mobility, ELLs don’t have the literacy skills expected by schools when they start schooling. They have to learn English and then their course materials, while undergoing social, emotional, and biological changes that come with immigration and going through puberty, for kids in middle-school and up.  

Most ELL students face a daunting achievement gap and do not perform as well as their native English speaking peers. Students do not succeed if they cannot read well. For example, only 6% of 4th grade ELLs read at or above proficiency compared with 36% non-ELL students. Also, only 12% of 4th grade ELLs scored at or above proficiency in maths compared with 41% non-ELLs. According to the Working Group on ELL Policy, academic performance in English and Math are tied to literacy skills as a high predictor of success, starting in middle school. Overall, ELLs must develop proficiency in English and knowledge of academic content in order to succeed in school, society, and later in the workforce.

COVID-19 disrupted the education of many students. Pre-COVID, there weren’t adequate resources, and the bilingual teacher shortage was widespread. And with the transition to remote learning, inequalities were exacerbated, as parents who don’t speak English most likely struggled to provide home support for English instruction to their kids, widening the learning gap between social class lines. Several students in lower-income families will have to work hard to catch up to their peers or be remediated. Our platform helps young ELL learners have access to quality and equitable learning environments, keeping them engaged and facilitating peer interaction, and guided independent work.

At the moment, we have a base prototype we are currently testing at a public school district with 52 schools, and two charter schools. 

MARVL builds on traditional vocabulary flashcards by adding animated AR “teachers” who coexist in the learner’s environment. Our animated characters provide robust, research-based best practices instruction – just as if a teacher were present. AR features add excitement and whimsy, keeping the child’s attention and increasing motivation. It also adds multimedia features that support learning, including animated demonstrations for verbs, sound effects, and animated text to emphasize letter-sound relationships.  The app engages learners and supports comprehension by providing examples in both English and the child’s home language. We also leverage AI feature for for adaptive, deep learning. 

At the moment, less than 1,000 users have/and are currently using our application. Due to COVID, we had to shift our distribution process to having parents voluntarily opt in, especially with the Manatee County schools. We are working with schools to get permission to go into schools to install our platform on their school computers. The good news is that both Dr. Smith and Ms. Olubela have been fully vaccinated, so we expect to be able to get into schools soon.

In a year, we project we project, we would have 100,000 users onboarded, as we are working on signing beta-testing agreements with schools and districts in other states. Our team often gets queries from teachers and superintendents interested in learning more about our product and possibly beta-testing it, some from abroad. We also have been fielding recommendations and introductions to prospective schools.

In the next five years, we project we would be a global company, operating at least in the U.S., Canada, Korea, and Nigeria. We project we would have onboarded at least 2.5M users in the U.S. and Canada, and at least 500,000 users in Korea and Nigeria. These numbers are conservatively estimated. 

We worked to ensure our goals are as SMART as possible. To achieve our goals, we would be breaking them into smaller tasks, tracked by a software like Metronome Growth System or the Entrepreneurial Operating System (we are browsing through different options at the moment), with KPIs assigned to each month, quarter, and year.

Some KPIs include monthly active user ratio, number of new contracts signed per period, number of engaged qualified leads in sales funnel, and customer satisfaction score. We would continually collect voluntarily provided data to gauge our user’s satisfaction, improve user interaction, deliver more value faster, get to the root of issues raised, and overall retain our users and turn them into life-long users.  

We would also be measuring our fundraising efforts, our brand awareness level, how many people and organizations we convinced to provide support to ELLs and minorities, and how many students participated in our mentorship program.

2 full time; 4 part times; and 6 contractors.

Our team is uniquely situated with research expertise and experiential knowledge relevant to this challenge, bringing perspectives as former students and educators. We have experienced being an English learner, developing multiple languages in childhood, and being second language international higher-edu students. We are:

Dr. Sara A. Smith, CEO: an Assistant Professor in Foreign Language and ESOL Education, she holds an MSc in Applied Linguistics and Second Language Education and a Ph.D. from the University of Oxford. MARVL was born from her research exploring technology enhanced solutions for ameliorating academic achievement inequities for linguistically diverse learners. Dr. Smith was a linguistically diverse elementary student and former ESOL teacher who taught across multiple contexts and grade levels

Murewa Olubela, MBA/MSEAT, COO: Ms. Olubela brings entrepreneurial and business expertise, having cofounded two startups before and worked as an intrapreneur at two tech startups. Ms. Olubela is Nigerian.

Dr. Maria Carlo: our Curriculum Head, holds a doctorate in Developmental Psychology from University of Massachusetts Amherst. She has experience creating bilingual curricula, and is focused on ensuring our materials are appropriate, user-friendly, effective, and adhere to pedagogical best practices. Dr. Carlo is a Spanish-English bilingual who attended schooling in the mainland U.S. and Puerto Rico. 

Dr. Sanghoon Park: our testing & analytics head, an Associate Professor of Learning Design and Technology with expertise in AR and learner motivation. He holds a doctorate in Instructional Systems from FSU and previously worked on multiple AR/VR tools for education. Dr. Park is a Korean-English bilingual.

As a women-founded and led company, with a BIPOC co-founder, we are fully committed to nurturing a diverse, equitable, and inclusive culture at MARVL. We vow that no one will be favored or treated unfairly because of who they are. And we understand that certain backgrounds start off with unearned advantages or disadvantages that perpetuate inequities in access, rewards, opportunity, and support.

For us, it’s more than just being an equal employment opportunity employer; it involves conscious deliberate actions that diversify the pools of applicants. Too often you hear people say they hired the best when put to task about diversity. But, that really is not true. They often hired the most promising in the limited pool they have, with certain unconscious biases coming into play.

They are multitudes of qualified candidates with diverse backgrounds, who for example could be our CTO, CFO, and Directors. We just have to make an extra effort to ensure they know about the positions open, and ensure our talent coordinator has a widened net when it comes to application. To this end, we are working on partnering with organizations committed to increasing the number of minorities employed in STEM positions.

We are applying in order to gain access to funding (grants and investments), a community of social impact leaders and peer support groups, mentorship, coaching and strategic advice from experts, and exposure in the media and at conferences. Becoming a Fellow would enable us to address and tackle our financial and technical expertise barriers.

We seek:

• Partners to participate in our fundraising rounds, either through investment and/or grants, and to help enhance our fundraising success.
• Experts in education and education-technology with experience in EdTech companies into schools to join our advisory board. 
• Marketing and Public Relations support to increase exposure and aid scaling.

• Partners such as Microsoft TEALS who support local STEM events and help schools develop and grow inclusive and sustainable STEM programs.
• Talent Search partners who work on ensuring leadership teams are inclusive and equitable.
• Organizations that work with refugees and refugee advocacy, a significant number of newcomer ELLs are refugees who face greater challenges acclimating academically. We would love to partner and provide our tech and resources to these kids. 

MARVL is a digital solution that provides evidence-based bilingual instruction and curriculum support for English language learners (ELLs) in K-12 schools, potentially reducing academic achievement inequalities. Users can use the solution at school or home, removing issues that teachers face when dealing with a limited class time and large class sizes. It works on any smartphone or tablet, is easily customized, can be aligned with existing curricula, and most of all, can be offered in many different languages.

A significant number of newcomers are refugee, who are mostly put directly into ESL classes, separated from other mainstream students. Since they come from country with unstable infrastructure, they often have interrupted learning. They might not even have attended school back home or learned basic skills in their own language. This affects their learning when they start schooling in the U.S., as mandated by USCIS. For example, a Somali student starting eight grade might not have learned to even write in Somali, facing an additional challenge when learning English and new vocabularies, significantly higher than an ELL student who has had formal education in their home language.

These kids have a limited amount of time to learn everything the rest of their peers have started learning since 1st grade. Winning the The Andan Prize for Innovation in Refugee Inclusion would mean being able to provide our technology for free for refugees, partner with organizations that cater to them, and being active in refugee communities. We have to engage the parents as the cultural background of their home countries are often different from the U.S., which expects parents to be active and somewhat responsible for the education of their children. Beyond helping parents to learn alongside their kids, forging familiar bonds and learning English themselves, we wish to help in any way we can. 

MARVL enables K-12 ELLs to grasp concepts taught in the classroom quickly and easily, especially within STEM classes. We are dedicated to making STEM education more accessible and equitable, for ELL students, including refugees. Linguistically diverse minorities are underrepresented in the U.S. STEM workforce, which can be traced to the interconnected nature of language skills and science and math instruction during early education. 

Our provision of an immersive and gamified bilingual learning environment, facilitates the mastery of Tier 2 and above vocabulary words, and increases student’s learning motivation and learning outcomes. With MARVL STEM teachers will be able to deliver robust, interactive, immersive, differentiated, academic vocabulary instruction with first language support to the ELLs in their classes.

Winning the GM Prize for Innovation in Refugee Inclusion prize enables us to continue our trajectory of making an impact, helping ELLs. We would use the money to continue to build our product and provide it freely for 15 months before commercializing it. 

Our solution leverages artificial intelligence and machine learning for adaptive, deep learning. We strive to create a technological solution that provides a learning experience an in-person bilingual teacher and a community of peers learning together offer. For learning to be long-lasting, it has to be conversational and immersive. AI enables us to provide an experience as close to in-person as possible. 

Research shows learner-specific differentiated instruction supported by AI models increases learning gains (Murphy, 2019). But AI models for language acquisition for emerging bilinguals and primary school aged children currently do not exist. We seek to fill in that gap. By merging AR with AI, we are able to provide an effective solution. AI and AR are cost-intensive technology. The The Patrick J. McGovern Foundation prize would aid our R&D efforts.

We endeavor to create a language acquisition AI solution, including a custom L2 AI model that provides new AI training models for language acquisition with a focus on K-12 students and emerging bilinguals/language learners, which currently do not exist. MARVL incorporates cognitive AI with vision recognition for text, allowing users to recognize words in texts. Our initial prototypes focused on traditional flashcards, and we are working on an updated version that includes support for print and object recognition, as well as speech and speaker recognition. Our speech and speaker recognition capacities will allow us to build models for language learner speech, with dynamic adaptive libraries, custom materials, and L1 specified materials. Our tech is supported by a pending patent.

For children from diverse backgrounds, English language proficiency is more strongly related to math achievement than for children from monolingual English-speaking homes (Foster et al, 2015; Foster et al., 2019). English  vocabulary in kindergarten, particularly academic vocabulary (for example, cognition words, scientific reasoning words) is essential for STEM success. Our technology will have broader, positive impacts on society by decreasing achievement inequities for children from linguistically diverse backgrounds and increasing representation in STEM fields. 

Through our customer discovery research, we learned that current classroom  solutions for ELLs are insufficient. Limited, minimally engaging software exists for children who  speak Spanish at home (such as Imagine Learning), there are no solutions for the 40% of families who speak other languages. District level stakeholders reported ESOL/bilingual teacher shortages and a  need for solutions for students who speak other languages (i.e. Arabic, Asian languages, Haitian Creole). Perspective customers reported that students do not enjoy current software programs, are demotivated, and learn slowly. Further, these solutions are not accessible to families without home  computers.  

Our innovation specifically addresses these needs. MARVL: 1) increases motivation and learning outcomes using AR and AI; 2) is available in multiple language combinations, and 3) is smartphone, tablet, and web compatible, thus accessible for families from different economic backgrounds.