NeurIPS

Connect with AI researchers shaping the future at NeurIPS workshops

NeurIPS workshops are where the real conversations happen, the ones that don't fit into a twenty-minute slot.

3 min readMachine Learning

If you're heading to NeurIPS this year and your calendar is already filling with workshops like WMHS, Physworld AI, or Vercodegen, the most valuable thing you can do right now is start the conversations before you board the plane. A Reddit user named navalsaras posted exactly that, a simple request to connect with fellow attendees ahead of time. It's a small ask, but it points to something bigger: the people shaping AI's future are often the ones who show up ready to talk, not just to listen.

This isn't just about networking. It's about recognizing that the real breakthroughs don't happen in the keynote halls. They happen in the hallways, over coffee, or in the workshop sessions where researchers dig into the specifics of their work. Our coverage of Reducing Dynamical Systems to One Parameter for Zero-Shot Reconstruction shows exactly the kind of focused, technical depth that defines NeurIPS workshops. That paper isn't a vague promise about the future, it's a concrete method for making complex systems interpretable and reconstructable. When you walk into a room where that kind of work is being presented, you want to be ready to ask the right questions. And the best way to prepare is to know who else is in that room.

At the same time, the logistics around NeurIPS have become a story of their own. We reported on how Area chair free passes prompt questions as NeurIPS sells out earlier this year, and the tension between access and exclusivity is only growing. When a conference sells out and passes are distributed unevenly, the workshops become even more important as spaces where the work, not the badge, matters. That's why a casual post like navalsaras's feels so timely. It cuts through the noise and says: I'm here to engage, not just to attend.

Our take is straightforward: if you're going to NeurIPS, don't treat it like a spectator event. The workshops are where the most interesting conversations live, and they reward preparation. Reach out to people beforehand. Read the papers that interest you, especially the ones that feel dense or niche. That ICLR 2027 Resets Its Scoring Scale for Paper Reviews article reminds us that the review process itself is constantly being refined, which means the bar for acceptance is shifting. The work that makes it into workshops like these has already passed a serious filter. Don't let that pass you by.

The specific takeaway: before you pack your bag, pick one workshop you care about most and find three people attending it. Send them a note. Ask about their work. That one step will change how you experience the conference, and it might change the direction of your own work, too.

From Machine Learning

Anyone going to WMHS, Physworld AI or Vercodegen? Would love to get to know some people beforehand!

Read the original at Machine Learning