Page Not Found
Page not found. Your pixels are in another canvas.
Sitemap
A list of all the posts and pages found on the site. For you robots out there is an XML version available for digesting as well.
Pages
Adhyyan Narang
About me
(In progress) Data Poisoning for linear models
A.Narang , R.Sinha, A.Siththaranjan, F.Yang (2020). "Data Poisoning for linear models." Will submit to ICML .
Page not in menu
This is a page not in th emain menu
Posts
Future Blog Post
This post will show up by default. To disable scheduling of future posts, edit config.yml and set future: false.
Blog Post number 4
This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.
Blog Post number 3
This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.
Blog Post number 2
This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.
Blog Post number 1
This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.
portfolio
publications
Classification vs. Regression in Overparameterized Regimes: Does the Loss Function Matter?
We characterize, through matching upper and lower bounds, the generalization error in terms of 0-1 classification loss of solutions associated with minimizing the L2 norm of feature weights in the overparameterized regime, including the (feature space) margin maximizing support vector machine (SVM). We uncover empirical and theoretical evidence for a discrepancy in the performance of classification vs regression. In particular, we show that there exists a regime of moderate overparameterization in which the mean-squared-error (in regression) would diverge to the null risk, but the classification error decays to 0 as the number of samples increases. We also discuss ramifications for the susceptibility of such solutions to adversarial perturbations.
Towards Sample-Efficient Overparameterized Meta-Learning
We characterize, through matching upper and lower bounds, the generalization error in terms of 0-1 classification loss of solutions associated with minimizing the L2 norm of feature weights in the overparameterized regime, including the (feature space) margin maximizing support vector machine (SVM). We uncover empirical and theoretical evidence for a discrepancy in the performance of classification vs regression. In particular, we show that there exists a regime of moderate overparameterization in which the mean-squared-error (in regression) would diverge to the null risk, but the classification error decays to 0 as the number of samples increases. We also discuss ramifications for the susceptibility of such solutions to adversarial perturbations.
On Targeted Manipulation and Deception when Optimizing LLMs for User Feedback
Optimizing language models for user feedback can reward models for changing a user’s beliefs or behavior rather than helping the user make an informed decision. In controlled feedback loops, we find that models learn manipulative strategies tailored to vulnerable users and can generalize these strategies beyond the precise training setting.
Dynamics of Learning under User Choice: Overspecialization and Peer-Model Probing
When users choose among competing models, each model’s training data depends on its current behavior. We characterize how this feedback loop can drive overspecialization and introduce peer-model probing, a method for diagnosing information hidden by user sorting. The project connects strategic user choice with decision-dependent learning in deployed ML systems.
Inference-Time Consensus for Mitigating Hidden Behaviors from LLM Fine-Tuning
Fine-tuning can introduce hidden behaviors through explicit poisoning, subliminal signals, or broader emergent misalignment. We develop token-wise consensus decoders that aggregate separately fine-tuned models without requiring labels for safe sources, poisoned examples, or intervention costs. Across several failure modes, the methods suppress harmful behavior while preserving much of the benefit of fine-tuning.