• Publications
  • Influence
Stolen Probability: A Structural Weakness of Neural Language Models
TLDR
This work presents numerical, theoretical and empirical analyses which show that words on the interior of the convex hull in the embedding space have their probability bounded by the probabilities of the word vectors on the hull.
Discovery of highly insecticidal synthetic spinosyn mimics - CAMD enabled de novo design simplifying a complex natural product.
Simplifying complex natural products: Computer modeling-based design leads to highly insecticidal, chemically simpler synthetic mimics of the spinosyn natural products that are active in the field. ©
De Novo Design of Potent, Insecticidal Synthetic Mimics of the Spinosyn Macrolide Natural Products
TLDR
Molecular modeling studies coupled with bioactivity-directed chemical modifications were used to define a less complex, synthetically accessible replacement for the spinosyn tetracycle, leading to the discovery of highly insecticidal analogs.
Controlling Global Statistics in Recurrent Neural Network Text Generation
TLDR
A dynamic regularizer that updates as training proceeds, based on the generative behavior of the RNNLMs is presented, which improves model perplexity when the statistical constraints are n-gram statistics taken from a large corpus.
Just Add Functions: A Neural-Symbolic Language Model
TLDR
A general methodology to enhance the inductive bias of NNLMs by incorporating simple functions into a neural architecture to form a hierarchical neural-symbolic language model (NSLM), which explicitly encode symbolic deterministic relationships to form probability distributions over words.
CODE: COMPILER-BASED NEURON-AWARE ENSEMBLE TRAINING
Deep Neural Networks (DNNs) are redefining the state-of-the-art performance in a variety of tasks like speech recognition and image classification. These impressive results are often enabled by
Who’s on First?: Probing the Learning and Representation Capabilities of Language Models on Deterministic Closed Domains
The capabilities of today’s natural language processing systems are typically evaluated using large datasets of curated questions and answers. While these are critical benchmarks of progress, they
Testing of the permeable Fe0-Barrier at dumpField Podlipa (Lubietova Cu-Deposit, Slovakia)
Material pochodzący z miejsca zrzutu wod kopalni Podlipa w opuszczonym zlozu Cu w Lubietovie wykazuje duzą zawartośc Cu, Fe, Cd, Co, Mn, Ni oraz Pb. W wiekszości przypadkow wyzsze zawartości metalu