Pooja Algikar (@algikarpooja) header
Pooja Algikar (@algikarpooja) profile picture

I love philosophy, all in for finding meaning in and of life with data-if there is any, PhD @vt_ece, postdoc @argonne Views are my own.

Start The MindLab Today
Joined November 2014
Photo by Pooja Algikar (@algikarpooja)Photo by Pooja Algikar (@algikarpooja)Photo by Pooja Algikar (@algikarpooja)Photo by Pooja Algikar (@algikarpooja)Photo by Pooja Algikar (@algikarpooja)Photo by Pooja Algikar (@algikarpooja)Photo by Pooja Algikar (@algikarpooja)Photo by Pooja Algikar (@algikarpooja)Photo by Pooja Algikar (@algikarpooja)Photo by Pooja Algikar (@algikarpooja)
24 citations in the first year. Not bad at all. Gaussian Process still got it. 1/n
6
1
99
12,382
I think I am a solar panel
7
143
I was just thinking there are like things happening to us all the time, and the only thing that makes them good or bad depends on our current state of “wanting” how crazy is that.
16
today's transfer-learning paper: Transferable Reinforcement Learning via Probabilistic Latent Embeddings and Dynamic Policy Adaptation for Sim-to-Real Deployment arxiv.org/abs/2605.27659
1
20
mbines a probabilistic context encoder, which estimates hidden environment properties from observed transitions, with distributional reinforcement learning, which estimates distributions of future rewards and costs. Its distinctive contribution is a deployment-time mechanism that connects uncertainty in the inferred environment to the conservatism of action selection. The result is a policy-transfer framework that can become more flexible as evidence about the target environment accumulates. 1/n In the paper's simulated autonomous-driving task, a vehicle controller encounters dynamics that differ from its training environments. The encoder processes recent states, actions, rewards, costs, and resulting states to infer a latent context describing those dynamics. A distributional cost critic then evaluates upper-tail cost quantiles, emphasizing potentially expensive outcomes when the context estimate is uncertain. The controller refines its proposed action using reward and cost gradients, with the risk level calibrated from simulations and adjusted as additional context becomes available. This builds on established probabilistic meta-RL and distributional RL methods; the new combination makes risk adjustment responsive to deployment uncertainty without updating the trained network parameters. 2/n
1
4
tional cost critic then evaluates upper-tail cost quantiles, emphasizing potentially expensive outcomes when the context estimate is uncertain. The controller refines its proposed action using reward and cost gradients, with the risk level calibrated from simulations and adjusted as additional context becomes available. This builds on established probabilistic meta-RL and distributional RL methods; the new combination makes risk adjustment responsive to deployment uncertainty without updating the trained network parameters. 2/n The authors evaluate the framework on simulated Safety Gym PointGoal2 navigation and autonomous driving, reporting improved reward-cost trade-offs under distribution shifts. In PointGoal2, the full method achieves a mean deployment cost of 8.86, compared with 10.12 when risk adaptation is disabled, against a cost threshold of 10. The ablation supports the value of coupling latent inference with adaptive risk control, although the added conservatism also reduces reward. The scientific insight is that uncertainty about the transferred environment can guide action-level adaptation after training. Its safety analysis assumes bounded context-estimation error and sufficient coverage of target dynamics by the training environments; the demonstrated results remain simulation-based, and modeled cost bounds do not establish physical-world safety. 3/n
3
Loving Vincent is absolutely stunning!
38
why are there always a lot of worst case scenario and only one best case scenario l?
1
76
Today's transfer-learning paper: Learning Domain Trajectories with Flow Matching for Gradual Domain Adaptation arxiv.org/abs/2610.30685
1
1
3
192
Illustrative example: A digit classifier trained on upright images must adapt to images rotated progressively through 15°, 30°, and 45°. FlowGDA learns local transport between successive domains while maintaining another connection to the labeled source representation. The local velocity field describes how features should move, while source-anchored transport and a semantic consistency loss encourage the identity of each digit to remain recognizable. Its two-stage sequential optimization stabilizes the learning process as adaptation progresses. Flow matching is an established technique; the paper's distinctive design is its combination with domain trajectories, source anchoring, and semantic preservation. 2/n
1
13
The authors report improvements over existing gradual domain adaptation methods, suggesting that learned continuous trajectories can be useful when distributions change substantially. The approach is particularly relevant when intermediate domains are meaningfully ordered and underlying class semantics remain reasonably stable. A central limitation is its dependence on a useful progression of domains and reliable semantic correspondence, leaving adaptation with emerging classes or unordered shifts less established. The theoretical and empirical scope should also be interpreted carefully: learning a continuous feature trajectory alone does not establish universal preservation of classification performance. Quantitative results and ablations require further verification because the full manuscript was inaccessible in this review. 3/n
5
Pooja Algikar retweeted
these are those last few days where you would flaunt sunshine!
1
1
12
448
Pooja Algikar retweeted
Reading Schrödinger, I find myself drawn to an almost Epicurean thought: curiosity may be the finest pleasure available to us. Something exists; I do not understand it; I want to know more. A simple sequence, and yet it can occupy a life. There is something quietly exhilarating about that thing. Everyone has lived with themselves all these years, and they still do not know the answer.
2
1
13
332
You think you do it for passion, for a goal, for whatever you hope will come of it. But maybe it doesn’t need to be for anything beyond the moment you’re in. Passion can fade; yesterday’s passion can become today’s act of will. And who knows, perhaps even the much-celebrated flow state is overrated. You can’t keep asking today to feel like yesterday. So do it for the present. Give your attention to what’s in front of you, even when the feeling that brought you there has changed. This is the part of your life you actually get to live.
64
You boast about consciousness, yet all you do is villare, because, though your mind works, your heart is darkened by depravity, and without a pure heart there can be no full, right consciousness. And how importunate you are, how you foist yourself, how you mug! Lies, lies, lies!" Dostoevsky
2
2
80
You can panic, plan, obsess, and overthink all you want, but eventually, you will realize that the only thing that truly matters is those moments spent sitting calmly, reading, and learning. The insights gained during these calm phases stay with you the longest; they shape your understanding, define your accumulated knowledge, and ultimately, they shape you.
1
91
Reading Schrödinger, I find myself drawn to an almost Epicurean thought: curiosity may be the finest pleasure available to us. Something exists; I do not understand it; I want to know more. A simple sequence, and yet it can occupy a life. There is something quietly exhilarating about that thing. Everyone has lived with themselves all these years, and they still do not know the answer.
2
1
13
332