Sian Gooding
@siangooding.bsky.social
Senior Research Scientist @GoogleDeepMind working on Autonomous Assistants ✍️🤖
You’ll collaborate with a kind, curious, research-driven team—including the brilliant @joao.omg.lol & @martinklissarov.bsky.social —and get to shape work at the frontier of multi-agent learning.
If that sounds like you, apply!
DM me if you're curious or have questions
If that sounds like you, apply!
DM me if you're curious or have questions
April 2, 2025 at 9:57 AM
You’ll collaborate with a kind, curious, research-driven team—including the brilliant @joao.omg.lol & @martinklissarov.bsky.social —and get to shape work at the frontier of multi-agent learning.
If that sounds like you, apply!
DM me if you're curious or have questions
If that sounds like you, apply!
DM me if you're curious or have questions
Some big questions we’re thinking about:
1⃣How do communication protocols emerge?
2⃣What inductive biases help coordination?
3⃣How can language improve generalisation and transfer?
1⃣How do communication protocols emerge?
2⃣What inductive biases help coordination?
3⃣How can language improve generalisation and transfer?
April 2, 2025 at 9:57 AM
Some big questions we’re thinking about:
1⃣How do communication protocols emerge?
2⃣What inductive biases help coordination?
3⃣How can language improve generalisation and transfer?
1⃣How do communication protocols emerge?
2⃣What inductive biases help coordination?
3⃣How can language improve generalisation and transfer?
We’re interested in:
🤖🤖 Multi-agent RL
🔠 Emergent language
🎲 Communication games
🧠 Social & cognitive modelling
📈 Scaling laws for coordination
🤖🤖 Multi-agent RL
🔠 Emergent language
🎲 Communication games
🧠 Social & cognitive modelling
📈 Scaling laws for coordination
April 2, 2025 at 9:57 AM
We’re interested in:
🤖🤖 Multi-agent RL
🔠 Emergent language
🎲 Communication games
🧠 Social & cognitive modelling
📈 Scaling laws for coordination
🤖🤖 Multi-agent RL
🔠 Emergent language
🎲 Communication games
🧠 Social & cognitive modelling
📈 Scaling laws for coordination
The project explores how agents can learn to communicate and coordinate in complex, open-ended environments—through emergent protocols, not hand-coded rules.
April 2, 2025 at 9:57 AM
The project explores how agents can learn to communicate and coordinate in complex, open-ended environments—through emergent protocols, not hand-coded rules.
Our full paper:
arxiv.org/pdf/2503.19711
arxiv.org/pdf/2503.19711
arxiv.org
April 2, 2025 at 9:51 AM
Our full paper:
arxiv.org/pdf/2503.19711
arxiv.org/pdf/2503.19711
Our work highlights the need for LLMs to improve in areas like action selection, self-evaluation + goal alignment to perform robustly in open-ended tasks
Implications of this work extend beyond writing assistance to autonomous workflows for LLMs in general open-ended use cases
Implications of this work extend beyond writing assistance to autonomous workflows for LLMs in general open-ended use cases
April 2, 2025 at 9:51 AM
Our work highlights the need for LLMs to improve in areas like action selection, self-evaluation + goal alignment to perform robustly in open-ended tasks
Implications of this work extend beyond writing assistance to autonomous workflows for LLMs in general open-ended use cases
Implications of this work extend beyond writing assistance to autonomous workflows for LLMs in general open-ended use cases
Finding: LLMs can lose track of the original goal during iterative refinement, leading to "semantic drift" - a divergence from the author's intent. This is a key challenge for autonomous revision. ✍️
April 2, 2025 at 9:51 AM
Finding: LLMs can lose track of the original goal during iterative refinement, leading to "semantic drift" - a divergence from the author's intent. This is a key challenge for autonomous revision. ✍️
Finding: LLMs struggle to reliably filter their own suggestions. They need better self-evaluation to work effectively in autonomous revision workflows. ⚖️
April 2, 2025 at 9:51 AM
Finding: LLMs struggle to reliably filter their own suggestions. They need better self-evaluation to work effectively in autonomous revision workflows. ⚖️
Finding: Gemini 1.5 Pro produced the highest quality editing suggestions, according to human evaluators, outperforming Claude 3.5 Sonnet and GPT-4o 🦾
April 2, 2025 at 9:51 AM
Finding: Gemini 1.5 Pro produced the highest quality editing suggestions, according to human evaluators, outperforming Claude 3.5 Sonnet and GPT-4o 🦾
Finding: LLMs tend to favour adding content, whereas human editors remove or restructure more. This suggests LLMs are sycophantic, reinforcing existing text rather than critically evaluating it. ➕
April 2, 2025 at 9:51 AM
Finding: LLMs tend to favour adding content, whereas human editors remove or restructure more. This suggests LLMs are sycophantic, reinforcing existing text rather than critically evaluating it. ➕
Why? There are many possible solutions and no single 'right' answer. Success is difficult to gauge!
We examine how LLMs generate + select text revisions, comparing their actions to human editors. We focus on action diversity, alignment with human prefs, and iterative improvement
We examine how LLMs generate + select text revisions, comparing their actions to human editors. We focus on action diversity, alignment with human prefs, and iterative improvement
April 2, 2025 at 9:51 AM
Why? There are many possible solutions and no single 'right' answer. Success is difficult to gauge!
We examine how LLMs generate + select text revisions, comparing their actions to human editors. We focus on action diversity, alignment with human prefs, and iterative improvement
We examine how LLMs generate + select text revisions, comparing their actions to human editors. We focus on action diversity, alignment with human prefs, and iterative improvement
Our paper explores this by analysing LLMs as autonomous co-writers. Work done with Lucia Lopez Rivilla, @egrefen.bsky.social ) 🫶
Open-ended tasks like writing are a real challenge for LLMs (even powerful ones like Gemini 1.5 Pro, Claude 3.5 Sonnet, and GPT-4o).
Open-ended tasks like writing are a real challenge for LLMs (even powerful ones like Gemini 1.5 Pro, Claude 3.5 Sonnet, and GPT-4o).
April 2, 2025 at 9:51 AM
Our paper explores this by analysing LLMs as autonomous co-writers. Work done with Lucia Lopez Rivilla, @egrefen.bsky.social ) 🫶
Open-ended tasks like writing are a real challenge for LLMs (even powerful ones like Gemini 1.5 Pro, Claude 3.5 Sonnet, and GPT-4o).
Open-ended tasks like writing are a real challenge for LLMs (even powerful ones like Gemini 1.5 Pro, Claude 3.5 Sonnet, and GPT-4o).