# forck.live — 18–24 July 2022 > AI lab announcements, news reports and podcasts. forck.live follows what companies announce about their own products. Press may corroborate an official story; it never leads and never appears as an item of its own. A missing date stays missing. A title is the publisher's words — where a company does not publish in English, an English rendering is listed separately as `Title in English` and named as this site's translation, never as theirs. This file is one issue of the weekly archive: every confirmed first-party announcement of ISO week 2022-W29, 18–24 July 2022, newest first. The window never moves, so this URL always names the same seven days — but a post published inside them and detected later is added when it arrives, so a closed week can still gain entries. Every entry is one first-party announcement, and carries the same eight fields in this order before any others: Company · Announced · Category · Coverage · Announcement · Group · Source · Record `Source` is the company's own page the announcement was traced to and `Record` is this site's page for it. Both are safe to cite. `Coverage` is a number of outlets, or `not counted` — which means this record reached no count of its own, never that nobody wrote about it. `Category` is this tracker's own analysis, and is authoritative wherever it is stated. `not stated` means either that nothing has analysed the post yet or that the analysis placed it outside this taxonomy; neither is a verdict about the announcement. `Announcement` is whether the post is news of something the company did, rather than a how-to, a customer story, a changelog listing or a staffing note. `Group` is which of this record's groups the entry falls in, carried on the entry itself so it holds where the groups are headings and where they are not. One of: `models`, `covered`, `announcements`, `routine`. Anything after `Record` is held detail that only some announcements carry. - This week's page: https://forck.live/week/2022-W29 - The current issue, and the catalogue of the other weeks: https://forck.live/briefing.md - Catalogue: https://forck.live/llms.txt ## Announcements ### LG AI Research — [CVPR 2022] Research Trend of Continual Learning - Company: LG AI Research (lgresearch.ai) - Announced: 2022-07-23 - Category: research-paper - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://www.lgresearch.ai/blog/view?seq=238 - Record: https://forck.live/items/4939-cvpr-2022-research-trend-of-continual-learning - Title in English (translated by forck.live from en): [CVPR 2022] Research Trend of Continual Learning - Subject: EXAONE This blog post from LG AI Research discusses recent research trends in continual learning presented at CVPR 2022, covering three directions: continual learning algorithms using Transformer models (DyTox), unsupervised/self-supervised continual learning, and proper application of batch normalization in continual learning. ### LG AI Research — [CVPR 2022] NeRF - Representing Scenes as Neural Radiance Fields for View Synthesis - Company: LG AI Research (lgresearch.ai) - Announced: 2022-07-22 - Category: research-paper - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://www.lgresearch.ai/blog/view?seq=237 - Record: https://forck.live/items/4940-cvpr-2022-nerf-representing-scenes-as-neural-radiance-fields-for-view-synthesis - Subject: EXAONE NeRF is a new method for view synthesis using neural radiance fields. ### Hugging Face — Advantage Actor Critic (A2C) - Company: Hugging Face (huggingface.co) - Announced: 2022-07-22 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://huggingface.co/blog/deep-rl-a2c - Record: https://forck.live/items/2175-advantage-actor-critic-a2c - Subject: Platform This is an educational article about the Advantage Actor Critic (A2C) algorithm, part of a deep reinforcement learning course. It explains the problem of variance in Reinforce and introduces Actor-Critic methods to reduce variance. ### Replit — Revamping the GitHub Import Flow - Company: Replit (replit.com) - Announced: 2022-07-21 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://replit.com/blog/nix-github-imports - Record: https://forck.live/items/17947-revamping-the-github-import-flow - Subject: Replit Agent Replit has updated its GitHub import flow to use Nix-based templates, replacing the previous Bash-style imports. The new flow automatically detects a repository's language via the GitHub API and recommends the correct template, preserving existing .replit and replit.nix configurations. The change aims to make imports faster and more intuitive, and brings Replit closer to fully migrating from its legacy Polygott backend. ### LG AI Research — LG AI Artist Tilda Winning the Award at New York Festivals - Company: LG AI Research (lgresearch.ai) - Announced: 2022-07-21 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://www.lgresearch.ai/news/view?seq=229 - Record: https://forck.live/items/11068-lg-ai-artist-tilda-winning-the-award-at-new-york-festivals - Subject: EXAONE LG AI Research's AI artist Tilda won gold and silver medals at the New York Festivals in the 'The Future Now: Best Use — Generative Creative' category for its work on climate change awareness and its identity book. ### LG AI Research — LG AI Artist Tilda Winning the Award at New York Festivals - Company: LG AI Research (lgresearch.ai) - Announced: 2022-07-21 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://www.lgresearch.ai/blog/view?seq=229 - Record: https://forck.live/items/4942-lg-ai-artist-tilda-winning-the-award-at-new-york-festivals - Subject: EXAONE LG AI Research's AI artist Tilda won gold and silver medals at the New York Festivals in the 'The Future Now: Best Use — Generative Creative' category for its work at New York Fashion Week and the Tilda Book. ### LG AI Research — [CVPR 2022] LG AI Research joins CVPR 2022, the Mecca of Vision Research - Company: LG AI Research (lgresearch.ai) - Announced: 2022-07-21 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://www.lgresearch.ai/blog/view?seq=236 - Record: https://forck.live/items/4941-cvpr-2022-lg-ai-research-joins-cvpr-2022-the-mecca-of-vision-research - Subject: EXAONE LG AI Research recaps its participation at CVPR 2022, highlighting research presentations including MERLOT Reserve, L-Verse, and Pix2NeRF. ### OpenAI — DALL·E now available in beta - Company: OpenAI (openai.com) - Announced: 2022-07-20T07:00:00+00:00 - Category: availability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://openai.com/index/dall-e-now-available-in-beta - Record: https://forck.live/items/898-dall-e-now-available-in-beta - Subject: GPT / ChatGPT / API OpenAI is making DALL·E available in beta, inviting 1 million people from the waitlist over the coming weeks. Users can create with DALL·E using free credits that refill every month, and can buy additional credits in 115-generation increments for $15. ### LG AI Research — [ICLR 2022] Part 3: Reinforcement Learning as a sequence modeling problem - Company: LG AI Research (lgresearch.ai) - Announced: 2022-07-20 - Category: research-paper - Coverage: not counted - Announcement: no - Group: routine - Source: https://www.lgresearch.ai/blog/view?seq=232 - Record: https://forck.live/items/4943-iclr-2022-part-3-reinforcement-learning-as-a-sequence-modeling-problem - Subject: EXAONE Decision Transformer (DT) and Generalized DT, which reformulate reinforcement learning as a sequence modeling problem using a transformer architecture, enabling stable policy learning via supervised learning without temporal difference learning. ### LG AI Research — [ICLR 2022] Part 2: Time Series Anomaly Detection - Company: LG AI Research (lgresearch.ai) - Announced: 2022-07-19 - Category: research-paper - Coverage: not counted - Announcement: no - Group: routine - Source: https://www.lgresearch.ai/blog/view?seq=231 - Record: https://forck.live/items/4944-iclr-2022-part-2-time-series-anomaly-detection - Subject: EXAONE Two papers on time series anomaly detection presented by LG AI Research at ICLR 2022, discussing deep learning-based methods and the Anomaly Transformer study. ### OpenAI — Reducing bias and improving safety in DALL·E 2 - Company: OpenAI (openai.com) - Announced: 2022-07-18T07:00:00+00:00 - Category: model-update - Coverage: not counted - Announcement: yes - Group: models - Source: https://openai.com/index/reducing-bias-and-improving-safety-in-dall-e-2 - Record: https://forck.live/items/899-reducing-bias-and-improving-safety-in-dall-e-2 - Subject: GPT / ChatGPT / API OpenAI is implementing a new technique in DALL·E to generate more diverse images of people, reducing bias and improving safety. ### LG AI Research — [ICLR 2022] Part 1: Efficient Transformer - Company: LG AI Research (lgresearch.ai) - Announced: 2022-07-18 - Category: research-paper - Coverage: not counted - Announcement: no - Group: routine - Source: https://www.lgresearch.ai/blog/view?seq=226 - Record: https://forck.live/items/4945-iclr-2022-part-1-efficient-transformer - Subject: EXAONE LG AI Research introduces three papers presented at ICLR 2022 on efficient Transformer implementation: Pyraformer, cosFormer, and PoNet, which aim to reduce the computational complexity of self-attention. ### Replicate — Exploring text to image models - Company: Replicate (replicate.com) - Announced: 2022-07-18 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://replicate.com/blog/exploring-text-to-image-models - Record: https://forck.live/items/2583-exploring-text-to-image-models - Subject: Platform To use their API to run text-to-image models, including GLID-3-XL, SRResCGAN, and ONGO, with code examples for image generation, enhancement, and variation.