Key Moments
Storied Navigation for Media Collections
Google TalksWant to know something specific about what's covered?
We've already dissected every moment. Ask and we will deliver (with timestamps).
Key Moments
MIT Media Lab's "media fabric" concept aims to create intelligent, story-driven navigation for vast digital archives, moving beyond simple lists and tags to meaningful engagement.
Key Insights
The MIT Media Lab's "media fabric" is envisioned as a semi-intelligent, evolving organism of media artifacts and structures that supports meaningful engagement with large collections.
The $100 laptop initiative, developed at the Media Lab, utilizes Linux, peer-to-peer mesh networks, and offline capabilities to maximize accessibility in low-connectivity regions.
Common sense reasoning, powered by OpenMind ConceptNet and LifeNet, is being explored to help computers understand and generate stories from media by parsing everyday activities and intentions.
Audio time-lapse, a project by an MIT student, aims to compress long audio streams by detecting and highlighting extraordinary events while blending them with ordinary ambiance.
Confectionary, a web-based media construction kit, offers a drag-and-drop interface for creators to assemble movies, images, and audio, encouraging content sharing via Creative Commons licenses.
The concept of 'storiedness' in media navigation addresses two key problems: intention (understanding the purpose behind media creation and consumption) and the elevation of ordinary events to an extraordinary status within a narrative.
The "media fabric" and its promise for digital archives
Glorianna Davenport introduces the concept of the "media fabric," a vision for a semi-intelligent, ever-evolving system designed to manage and organize the ever-growing volume of digital media. This fabric aims to move beyond simple archival lists and tags, offering a more meaningful and engaging way for users to interact with vast collections of audio and visual information. The core idea is to create structures and programs that support our engagement with media in ways that are personally relevant, tapping into creative story potential. This approach is particularly crucial given the exponential growth of personal and institutional media archives, which current interfaces often fail to navigate effectively.
Challenges with current media archiving and retrieval
Davenport highlights the limitations of existing archival systems, including institutional archives and even platforms like Flickr and YouTube, which often present media in overwhelming, linear lists or require highly specific, often limiting, search queries. She contrasts this with her personal experience accumulating thousands of images and hours of video, noting that a single image or video clip rarely tells a complete story. The overhead of manually sifting through vast personal collections to find coherent narratives is immense. Commercial venues, she notes, do not yet support the kind of deep, topic-coherent storytelling within personal archives that she envisions. The inability to easily extract and combine relevant segments from different media sources further hampers narrative construction.
Theoretical frameworks for "storiedness": intention and the extraordinary
Two key theoretical concepts underpin the idea of "storiedness": intention and the extraordinary. The problem of intention acknowledges that media is collected with individual purposes, which can conflict or evolve. Understanding the intent behind media creation and the evolving intent of the consumer is crucial for effective navigation. Davenport prefers to move beyond the limiting term 'metadata' to capture these nuances. The second concept, the extraordinary, posits that all stories, historically, have elevated moments of ordinary life to a special status. Whether it's a hero's journey or a child recounting a significant event, stories highlight what is unusual or impactful within the everyday. This notion of the "extraordinary event" helps in understanding the world and can be applied to making media collections more comprehensible and engaging.
Leveraging common sense reasoning for media analysis and creation
Davenport discusses projects employing "common sense reasoning" to imbue computers with a basic understanding of everyday life, crucial for media interpretation and story generation. Initiatives like "Open Mind Common Sense" crowdsourced common sense knowledge from the web, creating a semantic network of concepts and relationships. This forms the basis for systems that can suggest what a user might shoot next (as explored in a documentary camera prototype) or propose actions within a narrative context. For instance, a "comic kit" could suggest plausible next steps in a story based on common sense, such as planting a seed after picking it up. This system learns from user input, creating a two-way exchange of knowledge.
Developing tools for rich media construction and personalized storytelling
To facilitate narrative creation, the Media Lab has developed tools like "Confectionary," a web-based drag-and-drop interface that allows users to combine various media types (movies, images, audio) into stories. This tool emphasizes Creative Commons licensing, enabling users to define how their media can be used and shared, though it acknowledges limitations in integrating content from major media corporations. The platform also supports mentorship, allowing users to build stories privately before public release. Another project, "StoryNet," is evolving to help people build and share narrative sequences, aiming to create emergent stories from collective contributions and potentially provide a more accessible interface for media organization, especially for personal collections.
Exploring audio time-lapse and object-based media histories
Beyond visual media, research is also delving into audio as a narrative source. One project involves creating "audio time-lapse," which compresses extensive audio recordings by identifying and highlighting extraordinary events amidst ambient sound. This is being explored through instrumenting objects, like a "smart bench" equipped with sensors, to record and potentially narrate its own history. The idea is that objects witness events, and these witnessed "transactions" could hold value, giving objects a form of history that users might engage with. The system aims to detect events, match them with signatures, and classify them to produce a condensed, event-focused audio stream.
The $100 laptop initiative and its impact on connected learning
The $100 laptop project, a significant Media Lab initiative, aims to make computing accessible globally. These laptops run on Linux, utilize peer-to-peer mesh networks for local collaboration, and are designed to function with limited internet connectivity. To overcome power constraints, initial designs included a hand crank, though newer energy generation methods are being explored. The emphasis on mesh networking and local sharing is a deliberate choice to reduce reliance on costly internet connectivity and foster peer-to-peer learning. This approach is critical for deploying technology in regions where traditional infrastructure is scarce, encouraging collaborative problem-solving and documentation among young users.
Future directions: intention, sequencing, and emergent narratives
Looking ahead, the Media Lab is focusing on robust architectures for collecting knowledge about user intention and goals, essential for advanced media sequencing. Future iterations of projects like StoryNet aim to go deeper into organizing and understanding the context of media. For the $100 laptop initiative, a key challenge is how to mentor users, particularly children, to move beyond existing ideas and foster innovation. The development of platforms that allow users to document their processes and discoveries could lead to emergent narratives and innovative solutions bubbling up from communities. The overarching goal remains to create systems that make media collections intuitively navigable and personally meaningful, fostering deeper engagement and collaborative storytelling.
Mentioned in This Episode
●Software & Apps
●Companies
●Organizations
●People Referenced
Common Questions
The media fabric is conceptualized as a semi-intelligent organism composed of a vast and evolving collection of media artifacts, structures, and programs. It aims to support meaningful engagement with media in everyday life, synergizing with imagination and communication networks.
Topics
Mentioned in this video
A research institution founded in 1985, focusing on bringing together print, film, and computing, and exploring various research areas including digital media, intelligent machines, and human-machine partnerships.
A research group within the MIT Media Lab led by Gloriana Davenport, focusing on the emergence of a 'media fabric' - an intelligent organism of media artifacts supporting meaningful engagement.
An initiative run by Andy Lipman associated with MIT Media Lab, providing access to IP and researchers.
Manufacturer of the camera used in Barbara Berry's project for common sense reasoning in filmmaking.
A media company mentioned in the context of media licensing and content creation, regarding the use of their sound files.
An initiative that offers an internship program for high school and college students interested in history and photography, focusing on archival methods.
A news organization discussed in relation to its early 'stringer' model for news gathering and the eventual structuring of its archive.
Pioneer in digital media and co-founder of the MIT Media Lab, known for his work in getting industry to recognize the potential of digital media and partnering on the 'one hundred dollar laptop' initiative.
Researcher at the MIT Media Lab working on augmented exoskeleton technology with the goal of enhancing human mobility.
The new director of the MIT Media Lab as of January, formerly at Tivoli, bringing a vision for human-machine partnerships.
Runs the Digital Life Consortium associated with the MIT Media Lab.
Researcher at MIT Media Lab working on intelligent robots with affect for everyday assistance.
Researcher at MIT Media Lab studying cognitive architectures and early childhood language acquisition through extensive video and audio data collection.
Pioneering researcher at the MIT Media Lab, known for creating Logo and focusing on children's learning and the 'one hundred dollar laptop' initiative.
Former faculty at MIT Media Lab who developed 'OpenMind', a project to crowdsource common sense knowledge.
A student at MIT Media Lab who worked on using common sense reasoning for documentary filmmaking, developing the 'StoryNet' project.
A student working with Gloriana Davenport on a project where the input is a story and the output is also a story, exploring how systems can generate narratives from collections.
Worked in Thailand with children on workshops related to the $100 laptop initiative, focusing on innovative problem-solving and mentoring.
An open-source operating system mentioned as running on the $100 laptop, facilitating peer-to-peer collaboration.
Software used for image editing, mentioned as part of the overhead involved in creating simple media like a card for a grandson.
An image hosting service where users can tag photos, discussed as a tool for finding images based on keywords but lacking deeper coherence or narrative potential.
A database of concepts and everyday life events collected from web users, used as a resource for common sense reasoning.
A resource where users could associate items, play true/false games, and make suggestions to build short scripts about human actions.
A project, going into its second phase, that allows people to build short stories (sequences) on the web, which can then be used to understand narrative structures.
A platform that, when combined with SketchUp, offers new ways to think about building media environments using user-friendly tools.
3D modeling software purchased by Google, used by an architecture firm in New Orleans and integrated with Google Earth for creating media environments.
More from GoogleTalksArchive
View all 51 summaries
58 minEverything is Miscellaneous
54 minStatistical Aspects of Data Mining (Stats 202) Day 7
45 minKey Phrase Indexing With Controlled Vocabularies
63 minMysteries of the Human Genome
Ask anything from this episode.
Save it, chat with it, and connect it to Claude or ChatGPT. Get cited answers from the actual content — and build your own knowledge base of every podcast and video you care about.
Get Started Free