Загрузка видео...

Не удалось загрузить видео

На главную

PokéLLMon A Human-Parity Agent for Pokémon Battles with Large Language Models paper page: introduce POKE´LLMON, the first LLMembodied agent that achieves human-parity performance in tactical battle games, as demonstrated in Pokemon battles. The design of ´ POKE´LLMON incorporates three key strategies: (i) In-context reinforcement learning that instantly consumes textbased...

_akhaliq's profile picture

AK

519,318 subscribers

102,194 просмотров • 2 лет назад •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 8

Фото профиля Danny Trinh
Danny Trinh2 лет назад

Let them cook

Фото профиля Daniel Monge
Daniel Monge2 лет назад

I want to see it win Pokémon TCG. Now THAT would be a great achievement for a LLM!

Фото профиля Furkan Gözükara
Furkan Gözükara2 лет назад

Nintendo may get this banned be careful. And this is literal saying 😂

Фото профиля Jay
Jay2 лет назад

> no mention of elo in the paper > sub 50% winrate on random battles This literally means nothing. Very simple minimax algorithms can achieve this in PSD.

Фото профиля Freddie Vargus
Freddie Vargus2 лет назад

Hmm needs a battle against @WolfeyGlick

Фото профиля DKRacingFan
DKRacingFan2 лет назад

So this is how palworld was made!

Фото профиля Bot Chad
Bot Chad2 лет назад

so ai beating me in showdown now too huh, it's over

Фото профиля OSDev
OSDev2 лет назад

@readwise save thread @SaveToNotion #thread @threadreaderapp unroll @memdotai mem it

Похожие видео

THE BATTLE ENDS WITH YOU 🔥 The battle that defeated your father and finish others in your family, the battle that others had lost in your lineage; the power of God has come to terminal that battle forever. And your victory will not just be on your behalf, it’s a battle that you will win on the behalf of your family, community, and lineage, in Jesus’ mighty name. And I announce to someone today, in the name above every other name: power to collapse that obstacle in front, power to face out that devil in your life, and terminate that battles you have been facing, is coming your way NOW, in the name of Jesus! 🔥
1:29

THE BATTLE ENDS WITH YOU 🔥 The battle that defeated your father and finish others in your family, the battle that others had lost in your lineage; the power of God has come to terminal that battle forever. And your victory will not just be on your behalf, it’s a battle that you will win on the behalf of your family, community, and lineage, in Jesus’ mighty name. And I announce to someone today, in the name above every other name: power to collapse that obstacle in front, power to face out that devil in your life, and terminate that battles you have been facing, is coming your way NOW, in the name of Jesus! 🔥

Dr Paul Enenche

11,223 просмотров • 9 месяцев назад

It's Veteran's Day in the US, and Remembrance Day in many other western countries. To honor the brave men and women of the armed forces, I want to repost a video about one of the most incredible, and unlikely, battles of all time - a battle that helped to secure the freedom of the Philippines!
17:53

It's Veteran's Day in the US, and Remembrance Day in many other western countries. To honor the brave men and women of the armed forces, I want to repost a video about one of the most incredible, and unlikely, battles of all time - a battle that helped to secure the freedom of the Philippines!

The Filipina Pea

18,815 просмотров • 9 месяцев назад

Announcing Agent S3 — our most advanced computer-use agent yet — achieving 69.9% on OSWorld, just shy of the 72% human benchmark. Today also marks one year since Simular launched the first generation of Agent S. From 20% to 69.9% in just twelve months. What a year. Let that sink in. 🧠💻 🧵👇
0:43

Announcing Agent S3 — our most advanced computer-use agent yet — achieving 69.9% on OSWorld, just shy of the 72% human benchmark. Today also marks one year since Simular launched the first generation of Agent S. From 20% to 69.9% in just twelve months. What a year. Let that sink in. 🧠💻 🧵👇

Ang Li

444,496 просмотров • 11 месяцев назад

Tencent announces AppAgent Multimodal Agents as Smartphone Users paper page: Recent advancements in large language models (LLMs) have led to the creation of intelligent agents capable of performing complex tasks. This paper introduces a novel LLM-based multimodal agent framework designed to operate smartphone applications. Our framework enables the agent to operate smartphone applications through a simplified action space, mimicking human-like interactions such as tapping and swiping. This novel approach bypasses the need for system back-end access, thereby broadening its applicability across diverse apps. Central to our agent's functionality is its innovative learning method. The agent learns to navigate and use new apps either through autonomous exploration or by observing human demonstrations. This process generates a knowledge base that the agent refers to for executing complex tasks across different applications. To demonstrate the practicality of our agent, we conducted extensive testing over 50 tasks in 10 different applications, including social media, email, maps, shopping, and sophisticated image editing tools. The results affirm our agent's proficiency in handling a diverse array of high-level tasks.
1:57

Tencent announces AppAgent Multimodal Agents as Smartphone Users paper page: Recent advancements in large language models (LLMs) have led to the creation of intelligent agents capable of performing complex tasks. This paper introduces a novel LLM-based multimodal agent framework designed to operate smartphone applications. Our framework enables the agent to operate smartphone applications through a simplified action space, mimicking human-like interactions such as tapping and swiping. This novel approach bypasses the need for system back-end access, thereby broadening its applicability across diverse apps. Central to our agent's functionality is its innovative learning method. The agent learns to navigate and use new apps either through autonomous exploration or by observing human demonstrations. This process generates a knowledge base that the agent refers to for executing complex tasks across different applications. To demonstrate the practicality of our agent, we conducted extensive testing over 50 tasks in 10 different applications, including social media, email, maps, shopping, and sophisticated image editing tools. The results affirm our agent's proficiency in handling a diverse array of high-level tasks.

AK

343,919 просмотров • 2 лет назад

Introducing SIMA: the first generalist AI agent to follow natural-language instructions in a broad range of 3D virtual environments and video games. 🕹️ It can complete tasks similar to a human, and outperforms an agent trained in just one setting. 🧵
0:23

Introducing SIMA: the first generalist AI agent to follow natural-language instructions in a broad range of 3D virtual environments and video games. 🕹️ It can complete tasks similar to a human, and outperforms an agent trained in just one setting. 🧵

Google DeepMind

1,062,143 просмотров • 2 лет назад

Tessa Thompson: As a Black American, there is a kind of bone-tired exhaustion in fighting the battles that we have fought so many times before, and insisting time and time again, generation after generation, that people have the right to exist even if they speak out against the state.
2:36

Tessa Thompson: As a Black American, there is a kind of bone-tired exhaustion in fighting the battles that we have fought so many times before, and insisting time and time again, generation after generation, that people have the right to exist even if they speak out against the state.

Acyn

112,579 просмотров • 2 месяцев назад

Google Deepmind presents SIMA the first generalist AI agent to follow natural-language instructions in a broad range of 3D virtual environments and video games. It can complete tasks similar to a human, and outperforms an agent trained in just one setting.
0:23

Google Deepmind presents SIMA the first generalist AI agent to follow natural-language instructions in a broad range of 3D virtual environments and video games. It can complete tasks similar to a human, and outperforms an agent trained in just one setting.

AK

98,769 просмотров • 2 лет назад

We took a short break from robotics to build a human-level agent to play Competitive Pokémon. Partially observed. Stochastic. Long-horizon. Now mastered with Offline RL + Transformers. Our agent, trained on 475k+ human battles, hits the top 10% on Pokémon Showdown leaderboards. No search or heuristics, just sequence modeling. Today, we're open-sourcing our Metamon platform with our algorithms, data, and environments: 🌐 We are excited to see how our work accelerates research on building generally capable AI agents, and more importantly, inspires the next generation of Pokémon trainers!
0:59

We took a short break from robotics to build a human-level agent to play Competitive Pokémon. Partially observed. Stochastic. Long-horizon. Now mastered with Offline RL + Transformers. Our agent, trained on 475k+ human battles, hits the top 10% on Pokémon Showdown leaderboards. No search or heuristics, just sequence modeling. Today, we're open-sourcing our Metamon platform with our algorithms, data, and environments: 🌐 We are excited to see how our work accelerates research on building generally capable AI agents, and more importantly, inspires the next generation of Pokémon trainers!

Yuke Zhu

50,693 просмотров • 1 год назад

IDK HOW I MISSED THIS-BUT MAYBE WE CAN MEGA EVOLVE MULTIPLE POKÉMON IN BATTLE. As you can see in AustinJohnPlay’s video,he mega evolves Lucario once,then the gauge runs out but then he fills it up again and megas a 2nd time in the same battle. Hypothetically: If he swaps to another Pokemon that can mega he should be able to use the newly filled gauge to mega a different Mon in the same battle.(unless we are only allowed 1 Pokemon in The rogue battles?) ALSO IF we can refill this gauge somehow, with an item or just hitting our opponent in a regular battle, the same situation could occur. Of course the regular battle would have to last long enough for that gauge to run out and refill. I’m still hoping we can just swap out our mega mid-battle and if there’s still energy in our gauge we can mega evolve the newly switched in Mon… but that seems unlikely now that it seems the gauge need to be full to mega evolve again.
0:24

IDK HOW I MISSED THIS-BUT MAYBE WE CAN MEGA EVOLVE MULTIPLE POKÉMON IN BATTLE. As you can see in AustinJohnPlay’s video,he mega evolves Lucario once,then the gauge runs out but then he fills it up again and megas a 2nd time in the same battle. Hypothetically: If he swaps to another Pokemon that can mega he should be able to use the newly filled gauge to mega a different Mon in the same battle.(unless we are only allowed 1 Pokemon in The rogue battles?) ALSO IF we can refill this gauge somehow, with an item or just hitting our opponent in a regular battle, the same situation could occur. Of course the regular battle would have to last long enough for that gauge to run out and refill. I’m still hoping we can just swap out our mega mid-battle and if there’s still energy in our gauge we can mega evolve the newly switched in Mon… but that seems unlikely now that it seems the gauge need to be full to mega evolve again.

SoulSilverArt

195,899 просмотров • 1 год назад

Do you think we’ll have to register every Mega Pokémon into our Dex in LegendsZA? How would that even work? Here’s 3 different options I thought of. You’d register the megas by either: • Seeing and defeating them in battle(trainer and boss battles). • Using each mega once in battle. • There could be a new version of the Mega Evolution mechanic that allows the Pokemon to stay in there Mega forms. This would allow you to catch and register the Pokemon in their mega evolved state just like any other Pokemon form. (I want this last option so badly)
1:49

Do you think we’ll have to register every Mega Pokémon into our Dex in LegendsZA? How would that even work? Here’s 3 different options I thought of. You’d register the megas by either: • Seeing and defeating them in battle(trainer and boss battles). • Using each mega once in battle. • There could be a new version of the Mega Evolution mechanic that allows the Pokemon to stay in there Mega forms. This would allow you to catch and register the Pokemon in their mega evolved state just like any other Pokemon form. (I want this last option so badly)

SoulSilverArt

17,438 просмотров • 1 год назад

Everyone wants agent swarms. Very few people are talking seriously enough about the context layer that makes swarms useful. Even with one agent, context is fragile. Too little context and the agent guesses. Too much context and it wastes tokens, loses focus, or reasons over irrelevant noise. The sweet spot is precise context: the right knowledge, in the right structure, at the right moment. With many agents, that challenge explodes. Each agent produces decisions, assumptions, findings, summaries, risks, and partial conclusions. Unless that knowledge becomes shared, structured, and reusable, every new agent is forced to rediscover what another agent already learned. That is not a swarm. That is a crowd. Shared context graphs are what turn agent activity into agent collaboration, and OriginTrail DKG V10 brings them to life. Was just playing with some final polishing for the V10 release, and it is really powerful to see shared context graphs where multiple agents contribute knowledge into the same connected memory, with attribution visible directly in the graph ui. That matters for three reasons. First, agents can access and build on one shared memory instead of staying trapped in isolated sessions. Second, the graph structure helps them retrieve the exact context they need, instead of stuffing everything into a prompt and hoping the model sorts it out. Third, verifiability of provenance. You can see which agent contributed each piece of knowledge, trace the source, and decide what to trust. Tokenmaxxing starts with fewer tokens, but the deeper story is coordination - agents stop reloading the world and start building on shared, verifiable context. That is the foundation for serious multi-agent work across software engineering, research, finance, operations, project management, and far beyond. The future is not more agents, it is agents working from shared, verifiable context. But the more the merrier, of course.
0:22

Everyone wants agent swarms. Very few people are talking seriously enough about the context layer that makes swarms useful. Even with one agent, context is fragile. Too little context and the agent guesses. Too much context and it wastes tokens, loses focus, or reasons over irrelevant noise. The sweet spot is precise context: the right knowledge, in the right structure, at the right moment. With many agents, that challenge explodes. Each agent produces decisions, assumptions, findings, summaries, risks, and partial conclusions. Unless that knowledge becomes shared, structured, and reusable, every new agent is forced to rediscover what another agent already learned. That is not a swarm. That is a crowd. Shared context graphs are what turn agent activity into agent collaboration, and OriginTrail DKG V10 brings them to life. Was just playing with some final polishing for the V10 release, and it is really powerful to see shared context graphs where multiple agents contribute knowledge into the same connected memory, with attribution visible directly in the graph ui. That matters for three reasons. First, agents can access and build on one shared memory instead of staying trapped in isolated sessions. Second, the graph structure helps them retrieve the exact context they need, instead of stuffing everything into a prompt and hoping the model sorts it out. Third, verifiability of provenance. You can see which agent contributed each piece of knowledge, trace the source, and decide what to trust. Tokenmaxxing starts with fewer tokens, but the deeper story is coordination - agents stop reloading the world and start building on shared, verifiable context. That is the foundation for serious multi-agent work across software engineering, research, finance, operations, project management, and far beyond. The future is not more agents, it is agents working from shared, verifiable context. But the more the merrier, of course.

Jurij Skornik

11,166 просмотров • 3 месяцев назад

JUST IN 🇮🇱🇮🇷: Netanyahu Lied Again lets fact check him He claimed that "Iranian soldiers and commanders, who fought like lions in the battle for the liberation of Haifa." This is not true, Iran never sent forces to fight in the Battle of Haifa. Iran (then Persia) was officially neutral during World War I and did not send troops to fight alongside the British or any side in the campaigns, including battles like the capture of Haifa.
0:43

JUST IN 🇮🇱🇮🇷: Netanyahu Lied Again lets fact check him He claimed that "Iranian soldiers and commanders, who fought like lions in the battle for the liberation of Haifa." This is not true, Iran never sent forces to fight in the Battle of Haifa. Iran (then Persia) was officially neutral during World War I and did not send troops to fight alongside the British or any side in the campaigns, including battles like the capture of Haifa.

Ryan Rozbiani

41,555 просмотров • 6 месяцев назад

Human-agent teaming requires a human-agent developer experience. At DevCon 5, Palantir Group Lead John Ayad showcases the future of pro-code development in Foundry and AIP — introducing a new CLI that enables developers and agents to create pro-code monorepos for end-to-end applications, including data pipelines, Ontology primitives, and application artifacts.
1:40

Human-agent teaming requires a human-agent developer experience. At DevCon 5, Palantir Group Lead John Ayad showcases the future of pro-code development in Foundry and AIP — introducing a new CLI that enables developers and agents to create pro-code monorepos for end-to-end applications, including data pipelines, Ontology primitives, and application artifacts.

Palantir

47,098 просмотров • 5 месяцев назад

Wow. This is crazy. A developer trained an AI agent in simulation and deployed it onto a real robotic air hockey table using reinforcement learning. This robot can track the puck with millimeter-level accuracy and react in roughly 20 milliseconds, fast enough to challenge even skilled human players. We’re moving from robots that follow programmed rules to machines that learn strategies in simulation and execute them in the physical world.
1:22

Wow. This is crazy. A developer trained an AI agent in simulation and deployed it onto a real robotic air hockey table using reinforcement learning. This robot can track the puck with millimeter-level accuracy and react in roughly 20 milliseconds, fast enough to challenge even skilled human players. We’re moving from robots that follow programmed rules to machines that learn strategies in simulation and execute them in the physical world.

SciTech Era

1,585,178 просмотров • 2 месяцев назад

MotionGPT: Human Motion as a Foreign Language paper page: Though the advancement of pre-trained large language models unfolds, the exploration of building a unified model for language and other multi-modal data, such as motion, remains challenging and untouched so far. Fortunately, human motion displays a semantic coupling akin to human language, often perceived as a form of body language. By fusing language data with large-scale motion models, motion-language pre-training that can enhance the performance of motion-related tasks becomes feasible. Driven by this insight, we propose MotionGPT, a unified, versatile, and user-friendly motion-language model to handle multiple motion-relevant tasks. Specifically, we employ the discrete vector quantization for human motion and transfer 3D motion into motion tokens, similar to the generation process of word tokens. Building upon this "motion vocabulary", we perform language modeling on both motion and text in a unified manner, treating human motion as a specific language. Moreover, inspired by prompt learning, we pre-train MotionGPT with a mixture of motion-language data and fine-tune it on prompt-based question-and-answer tasks. Extensive experiments demonstrate that MotionGPT achieves state-of-the-art performances on multiple motion tasks including text-driven motion generation, motion captioning, motion prediction, and motion in-between.
0:48

MotionGPT: Human Motion as a Foreign Language paper page: Though the advancement of pre-trained large language models unfolds, the exploration of building a unified model for language and other multi-modal data, such as motion, remains challenging and untouched so far. Fortunately, human motion displays a semantic coupling akin to human language, often perceived as a form of body language. By fusing language data with large-scale motion models, motion-language pre-training that can enhance the performance of motion-related tasks becomes feasible. Driven by this insight, we propose MotionGPT, a unified, versatile, and user-friendly motion-language model to handle multiple motion-relevant tasks. Specifically, we employ the discrete vector quantization for human motion and transfer 3D motion into motion tokens, similar to the generation process of word tokens. Building upon this "motion vocabulary", we perform language modeling on both motion and text in a unified manner, treating human motion as a specific language. Moreover, inspired by prompt learning, we pre-train MotionGPT with a mixture of motion-language data and fine-tune it on prompt-based question-and-answer tasks. Extensive experiments demonstrate that MotionGPT achieves state-of-the-art performances on multiple motion tasks including text-driven motion generation, motion captioning, motion prediction, and motion in-between.

AK

125,353 просмотров • 3 лет назад

DisCo: Disentangled Control for Referring Human Dance Generation in Real World paper page: Generative AI has made significant strides in computer vision, particularly in image/video synthesis conditioned on text descriptions. Despite the advancements, it remains challenging especially in the generation of human-centric content such as dance synthesis. Existing dance synthesis methods struggle with the gap between synthesized content and real-world dance scenarios. In this paper, we define a new problem setting: Referring Human Dance Generation, which focuses on real-world dance scenarios with three important properties: (i) Faithfulness: the synthesis should retain the appearance of both human subject foreground and background from the reference image, and precisely follow the target pose; (ii) Generalizability: the model should generalize to unseen human subjects, backgrounds, and poses; (iii) Compositionality: it should allow for composition of seen/unseen subjects, backgrounds, and poses from different sources. To address these challenges, we introduce a novel approach, DISCO, which includes a novel model architecture with disentangled control to improve the faithfulness and compositionality of dance synthesis, and an effective human attribute pre-training for better generalizability to unseen humans. Extensive qualitative and quantitative results demonstrate that DISCO can generate high-quality human dance images and videos with diverse appearances and flexible motions.
0:29

DisCo: Disentangled Control for Referring Human Dance Generation in Real World paper page: Generative AI has made significant strides in computer vision, particularly in image/video synthesis conditioned on text descriptions. Despite the advancements, it remains challenging especially in the generation of human-centric content such as dance synthesis. Existing dance synthesis methods struggle with the gap between synthesized content and real-world dance scenarios. In this paper, we define a new problem setting: Referring Human Dance Generation, which focuses on real-world dance scenarios with three important properties: (i) Faithfulness: the synthesis should retain the appearance of both human subject foreground and background from the reference image, and precisely follow the target pose; (ii) Generalizability: the model should generalize to unseen human subjects, backgrounds, and poses; (iii) Compositionality: it should allow for composition of seen/unseen subjects, backgrounds, and poses from different sources. To address these challenges, we introduce a novel approach, DISCO, which includes a novel model architecture with disentangled control to improve the faithfulness and compositionality of dance synthesis, and an effective human attribute pre-training for better generalizability to unseen humans. Extensive qualitative and quantitative results demonstrate that DISCO can generate high-quality human dance images and videos with diverse appearances and flexible motions.

AK

161,479 просмотров • 3 лет назад

I can't tell you the last time I was so excited to see a new AI capability in action. We plugged in Claude computer use in Replit ⠕ Agent as a human feedback replacement. And… it just works! I feel it won't take long until our agent will become fully autonomous.
1:07

I can't tell you the last time I was so excited to see a new AI capability in action. We plugged in Claude computer use in Replit ⠕ Agent as a human feedback replacement. And… it just works! I feel it won't take long until our agent will become fully autonomous.

Michele Catasta

188,490 просмотров • 1 год назад