Continue Reading

Using Thumbnail Montages To Optimize AI-Based OCR Speed & Costs: Part 6 – Further Grid Layout Experiments

Yesterday's experiments on how montage layouts impact OCR accuracy suggested that there were few differences between vertical, horizontal and grid-based…

Continue Reading

Independent: No, Young Black Men Aren't To Blame If Trump Wins The Election

A fascinating analysis of the 2024 US presidential election by Richard Addy and Luba Kassova. Read The Full Article.

Continue Reading

Using Thumbnail Montages To Optimize AI-Based OCR Speed & Costs: Part 5 – The Impact Of Grid Layout

Continuing our OCR experiments, how much of an impact does the layout of the montage grid have on OCR performance?…

Continue Reading

Using Thumbnail Montages To Optimize AI-Based OCR Speed & Costs: Part 4

Continuing our experiments with optimizing video OCR costs, what impact does image resolution have on OCR accuracy? Assuming we are…

Continue Reading

Using Thumbnail Montages To Optimize AI-Based OCR Speed & Costs: Part 3

Continuing our experiments in using montages to optimize OCR speed and cost for the TV News Archive, what might it…

Continue Reading

Using Thumbnail Montages To Optimize AI-Based OCR Speed & Costs: Part 2

Continuing yesterday's experiments, what if we boost the resolution of the surrogate thumbnail images from 200×200 pixels to 500×500 pixels…

Continue Reading

Lessons From The Trenches: How OCR Text Search Beat Out Visual Search For Television News

Just over six and a half years ago, in collaboration with the Internet Archive's TV News Archive, we launched the…

Continue Reading

Using Thumbnail Montages To Optimize AI-Based OCR Speed & Costs: Part 1

One of the most fascinating lessons we've learned from working with journalists and scholars using the Internet Archive's TV News…

Continue Reading

Replacing The Current Single-Frame Broadcast Thumbnails With Algorithmically-Selected Montage Thumbnails: Part 2

How might replacing our current single-frame television news broadcast thumbnails with montage thumbnails of algorithmically-selected representative frames from throughout a…

Continue Reading

Replacing The Current Single-Frame Broadcast Thumbnails With Algorithmically-Selected Montage Thumbnails

Instead of displaying a single isolated frame as the thumbnail for each television news broadcast in the Visual Explorer, what…

Continue Reading

Using Algorithmic Thumbnail Selection To Create Montage Preview Thumbnails For Television News

Continuing our series on algorithmic thumbnail selection for visual summarization of television news broadcasts, perhaps the most important question is…

Continue Reading

La Silla Vacia: Aidiosyncrasy And Memocracy

A fascinating comparison of media tone and approval ratings in Columbia. Read The Full Article.

Continue Reading

HUMVI: A Multilingual Dataset for Detecting Violent Incidents Impacting Humanitarian Aid

A paper by researchers at Dataminr, Insecurity Insight and Northwestern University uses GDELT as a global media index. Read The…

Continue Reading

More Experiments In Algorithmic Thumbnail Generation For Television News: Evening News Broadcasts

Continuing our series on algorithmic thumbnail generation for television news, what about a few examples drawn from evening news broadcasts,…

Continue Reading

Using Algorithmic Thumbnail Generation To Yield Better Single-Image Thumbnail Representations

Over the past several days we've explored applying ffmpeg's built-in representative frame detection "thumbnail" filter, returning to our earliest experiments…

Continue Reading

ChatGPT's o1-Preview "Reasoning Model" Fairs Little Better Than 4o On Our Thumbnail Experiments

Yesterday we explored the ability of ChatGPT's 4o foundational LLM to act as a brainstorming partner in the development of…

Continue Reading

ChatGPT's Unhelpful Take On Visually Summarizing Television News Broadcasts Through Thumbnails

Continuing our series on selecting "representative" frames from television news broadcasts to visually summarize them, thus far we've been experimenting…

Continue Reading

Visually Summarizing Television News Broadcasts Through Thumbnails: Algorithmic Vs Time-Based Experiments: Part 2

Continuing our experiments from yesterday creating representative thumbnails of a television news broadcast, let's look at how the two approaches…

Continue Reading

Scaling New Heights: Transformative Cross-GPU Sampling for Training Billion-Edge Graphs

A new paper by researchers at Wuhan University, NVIDIA and the University of Macau: Efficient training of Graph Neural Networks…

Continue Reading

Visually Summarizing Television News Broadcasts Through Thumbnails: Algorithmic Vs Time-Based Experiments

The Visual Explorer visually summarizes television news broadcasts through fixed 1/4fps thumbnail grids – an approach developed through extensive human…

Continue Reading

Comparing Television News Coverage Of Trump Vs Kamala Vs Biden

How are the two present and one former presidential candidates being covered on television news? The timeline below compares their…

Continue Reading

A Digital Twin Glimpse At The Internet Archive's TV News Archive: 27 Billion Seconds Over 10.9 Million Broadcasts From 327 Channels In 50+ Countries & 150+ Languages Over A Quarter-Century

Using our new BigQuery + Bigtable GCS digital twin, we can look across the entire Internet Archive's TV News Archive…

Continue Reading

Behind The Scenes: Network Intelligence Topology Mapping By GCP Service

Last year we explored how GCP's Network Intelligence Network Topology mapping can be used to understand the network flows across…

Continue Reading

WashPost: No Amount Of Evidence Will Convince Republicans Of Trump's 2020 Guilt

The Washington Post's Philip Bump examines media coverage of Donald Trump. Read The Full Article.