Zobrazeno 1 - 10
of 28
pro vyhledávání: '"Rajendran, Goutham"'
Large Language Models (LLMs) have the capacity to store and recall facts. Through experimentation with open-source models, we observe that this ability to retrieve facts can be easily manipulated by changing contexts, even without altering their fact
Externí odkaz:
http://arxiv.org/abs/2406.18400
Autor:
Bakshi, Ainesh, Kothari, Pravesh, Rajendran, Goutham, Tulsiani, Madhur, Vijayaraghavan, Aravindan
A set of high dimensional points $X=\{x_1, x_2,\ldots, x_n\} \subset R^d$ in isotropic position is said to be $\delta$-anti concentrated if for every direction $v$, the fraction of points in $X$ satisfying $|\langle x_i,v \rangle |\leq \delta$ is at
Externí odkaz:
http://arxiv.org/abs/2405.15084
Recent works have argued that high-level semantic concepts are encoded "linearly" in the representation space of large language models. In this work, we study the origins of such linear representations. To that end, we introduce a simple latent varia
Externí odkaz:
http://arxiv.org/abs/2403.03867
To build intelligent machine learning systems, there are two broad approaches. One approach is to build inherently interpretable models, as endeavored by the growing field of causal representation learning. The other approach is to build highly-perfo
Externí odkaz:
http://arxiv.org/abs/2402.09236
We investigate the relationship between system identification and intervention design in dynamical systems. While previous research demonstrated how identifiable representation learning methods, such as Independent Component Analysis (ICA), can revea
Externí odkaz:
http://arxiv.org/abs/2311.18048
Autor:
Buchholz, Simon, Rajendran, Goutham, Rosenfeld, Elan, Aragam, Bryon, Schölkopf, Bernhard, Ravikumar, Pradeep
We study the problem of learning causal representations from unknown, latent interventions in a general setting, where the latent distribution is Gaussian but the mixing function is completely general. We prove strong identifiability results given un
Externí odkaz:
http://arxiv.org/abs/2306.02235
Given a graph and an integer $k$, Densest $k$-Subgraph is the algorithmic task of finding the subgraph on $k$ vertices with the maximum number of edges. This is a fundamental problem that has been subject to intense study for decades, with applicatio
Externí odkaz:
http://arxiv.org/abs/2303.17506
Autor:
Rajendran, Goutham
We develop new tools in the theory of nonlinear random matrices and apply them to study the performance of the Sum of Squares (SoS) hierarchy on average-case problems. The SoS hierarchy is a powerful optimization technique that has achieved tremendou
Externí odkaz:
http://arxiv.org/abs/2302.04462
Autor:
Rajendran, Goutham, Tulsiani, Madhur
Analyzing concentration of large random matrices is a common task in a wide variety of fields. Given independent random variables, many tools are available to analyze random matrices whose entries are linear in the variables, e.g. the matrix-Bernstei
Externí odkaz:
http://arxiv.org/abs/2209.02655
Autor:
Rajendran, Goutham, Zou, Wei
We investigate robustness properties of pre-trained neural models for automatic speech recognition. Real life data in machine learning is usually very noisy and almost never clean, which can be attributed to various factors depending on the domain, e
Externí odkaz:
http://arxiv.org/abs/2208.08509