A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.
The Gumbel-max trick and its extensions simplify sampling from categorical distributions in machine learning.
Proposes a new method for estimating counterfactual treatment effects.
The well-known Gumbel-Max Trick for sampling elements from a categorical distribution (or more generally a nonnegative vector) and its variants have been widely used in areas such as machine learning and information retrieval. To sample a random element (or a Gumbel-Max variable ) in proportion to its positive w…
Study on estimating Gumbel--Max watermark proportions in edited documents.
Estimates proportions of LLM-generated text in mixed documents.
Unified framework for gradient estimation in combinatorial spaces.
The paper develops a method for inferring second opinions from experts using counterfactual inference.
Reparameterization of variational auto-encoders with continuous random variables is an effective method for reducing the variance of their gradient estimates. In the discrete case, one can perform reparametrization using the Gumbel-Max trick, but the resulting objective relies on an operation and is non-dif…
Thompson sampling has impressive empirical performance for many multi-armed bandit problems. But current algorithms for Thompson sampling only work for the case of conjugate priors since these algorithms require to infer the posterior, which is often computationally intractable when the prior is not conjugate. In this …
We introduce an off-policy evaluation procedure for highlighting episodes where applying a reinforcement learned (RL) policy is likely to have produced a substantially different outcome than the observed policy. In particular, we introduce a class of structural causal models (SCMs) for generating counterfactual traject…
Many machine learning tasks require sampling a subset of items from a collection based on a parameterized distribution. The Gumbel-softmax trick can be used to sample a single item, and allows for low-variance reparameterized gradients with respect to the parameters of the underlying distribution. However, stochastic o…
The well-known Gumbel-Max trick for sampling from a categorical distribution can be extended to sample elements without replacement. We show how to implicitly apply this 'Gumbel-Top-' trick on a factorized distribution over sequences, allowing to draw exact samples without replacement using a Stochastic Beam Sea…
Log-linear models are arguably the most successful class of graphical models for large-scale applications because of their simplicity and tractability. Learning and inference with these models require calculating the partition function, which is a major bottleneck and intractable for large state spaces. Importance Samp…
Develops methods to answer counterfactual questions in temporal point processes.
New framework improves text watermark detection under imperfect pseudorandomness.
Develops methods for finding counterfactual explanations in sequential decision making.
LLM-as-a-service prices vary arbitrarily due to tokenization multiplicity.
New method detects watermarks in LLM-generated text with human edits.