Peace, no war 🔫
The Belamy | Top AI stories of last week
 |
The ongoing Ukraine-Russia tussle seems to have shaken the world in unimaginable ways, but the tech community stands strong and united. The likes of Grammarly, Telegram, GitLab, etc. originated from Ukraine, alongside tech giants like Google, Apple, Meta (formerly Facebook), Uber and Twitter are all showering their supportive gestures.
Grammarly went blue yellow for Ukraine, alongside announcing to donate $5 million to businesses and funds supporting the people of Ukraine. Google contributed $15 million in donations to aid relief efforts in Ukraine.
Apple, on the other hand, paused sales of its products to Russia, and removed Russian state-funded news publishers such as RT News and Sputnik News from its app store, alongside disabling traffic updates for Ukraine in Apple Maps and others.
Meanwhile, Meta announced that it is prohibiting Russian state media from running ads or monetising on its platform anywhere in the world. Uber offered free rides to the Polish border, transporting food and essential goods, funding humanitarian aid, and launching in-app donations to support relief efforts. Airbnb pledged to provide free, short-term housing for up to 100,000 refugees.
The Tech to the Rescue Foundation recently launched a campaign called #TechForUkraine , which is pledged to support the Ukrainian people. More than 225 companies in over 25 countries and 300 individuals have pledged their support. This includes Amazon Web Services, Microsoft, Salesforce, and others.
Read the tech leaders' reaction to the Ukraine Russia-conflict here.
Startups, Apps & Tools
Based in Bengaluru, Jiffy.ai is a software company that offers AI-powered intelligent and integrated platforms for the digital enterprise. From processing unstructured data that is captured in different formats (images, pics, videos, etc.) to make cognitive decisions, and using NLP and ML-based algorithms, Jiffy claims to have helped customers achieve close to 90 percent automation.
San Francisco-based AI-enabled video editing tool Descript builds simple and powerful collaborative tools for new media creators: podcasters, journalists, YouTubers, and more. The platform allows you to edit audio or video simply by editing a text transcript – cut, paste, and delete – and also it edits the audio or video to match.
At the Inside the Lab event, Meta introduced TorchRec, an open-source library for building SOTA recommendations systems under PyTorch. This new library provides common sparsity and parallelism primitives, enabling researchers to build personalisation models and deploy them in production. It includes a scalable low-level modelling foundation alongside rich batteries-included modules.
Read more about TorchRec here.
AIM Videos
Want to know how Spotify Spots your favourite song? Check out the video below.
Research & Papers
Bayesian Structure Learning with Generative Flow Networks
Last week, Bengio & co published Bayesian structure learning with generative flow networks. The team proposed using a generative flow network (GFlowNet) as an alternative to MCMC for approximating the posterior distribution over the structure of Bayesian networks, given a dataset of observations. In this research, the team evaluated simulated and real data and showed a new approach called DAG-GFlowNet, which provided an accurate approximation of the posterior over DAGs, and compared well against other methods based on MCMC or variational inference.
Label-free Explainability for Unsupervised Models
Researchers from University of Cambridge – Jonathan Crabbé , Mihaela van der Schaar – introduced two crucial extensions of post-hoc explanation techniques like label-free feature importance and label-free example importance that highlight influential features and training examples for a black-box to construct representations at inference time. In this research paper, they demonstrated that their extensions can be successfully implemented as simple wrappers around many existing feature and example importance methods. They also illustrated the utility of their label-free explainability paradigm through a qualitative and quantitative comparison of representation spaces learned by various autoencoders trained on distinct unsupervised tasks.
Social media dialogues
Article in Focus
AIM Long Reads: Best Indian Cities to Launch AI startups
The Indian tech startup space exploded in 2021 with over 2,250 tech startups and 42 unicorns added in a year’s span. India is now the third-largest startup ecosystem in the world after the US and China. According to the Economic Survey 2021-22, 555 districts in India had at least one startup. While most founders prefer Tier I cities to launch their startups, the pandemic has brought a paradigm shift as companies have moved to a remote or hybrid working model.
We have parsed scores of data and research and sounded out key stakeholders to understand if the location is an X factor in the success of a startup. We looked at parameters such as infrastructure, state policies, quality of life, educational institutions and talent pool etc to arrive at our outcomes.
From a numbers perspective, Bengaluru is leading the way with 1,064 AI startups , followed by Mumbai (363), Delhi (317) and Hyderabad (287). Bengaluru has close to 256 pure-play AI and analytics startups, followed by Delhi NCR (190), Mumbai (120), Pune (78), Hyderabad (74), Chennai (53), and Ahmedabad (26).
Top News of the Week
83 per cent of Indian BFSI Players Think AI is Compulsory for Better Customer Experience
3M to Open Tech Hub in Bengaluru, Plans to Hire 1,000 Employees
Intel and AMD Halt Sales of Industrial Chips to Russia
Create your profile
Only paid subscribers can comment on this post
Check your email
For your security, we need to re-authenticate you.
Click the link we sent to , or click here to sign in.