We introduce a method to stabilize Generative Adversarial Networks (GANs) by defining the generator objective with respect to an unrolled optimization of the discriminator. This allows training to be adjusted between using the optimal discriminator in the generator’s objective, which is ideal but infeasible in practice, and using the current value of the discriminator, which is often unstable a...

An attempt is being made to construct an FM broadcast receiver utilizing relatively few tubes but incorporating the principles of multipath interference reduction. The parts to be constructed are the i-f amplifier, a two-tube limiter, a discriminator and an audio amplifier stage. A bandwidth of 2 Mc for the limiter and discriminator was chosen after a study of multipath interference in cities s...

For each clone C on a set A there is an associated equivalence relation, called C-equivalence, on the set of all operations on A, which relates two operations iff each one is a substitution instance of the other using operations from C. In this paper we prove that if C is a discriminator clone on a finite set, then there are only finitely many C-equivalence classes. Moreover, we show that the s...

In this paper we introduce the notion of minimum-weight edge-discriminators in hypergraphs, and study their various properties. For a hypergraph H = (V,E ), a function λ : V → Z+∪{0} is said to be an edge-discriminator onH if∑v∈Ei λ(v) > 0, for all hyperedges Ei ∈ E , and ∑ v∈Ei λ(v) 6= ∑ v∈Ej λ(v), for every two distinct hyperedges Ei, Ej ∈ E . An optimal edge-discriminator on H, to be denoted...

A new unambiguous discriminator similar to a conventional Double Delta correlator is tailored for sinephased BOC(1,1) signal tracking. It is shown in this paper that it has efficient multipath mitigation at the cost of degraded noise resistance due to correlation loss of using waveform subtraction. Its multipath performances is evaluated in both coherent and dot-product type noncoherent structu...

Query auto-completion (QAC) serves as a critical functionality in contemporary textual search systems by generating real-time query completion suggestions based on user’s input prefix. Despite the prevalent use of language models (LMs) QAC candidate generation, LM-based approaches frequently suffer from overcorrection issues during pair-wise loss training and efficiency deficiencies. To address...

Preface Aknowledgements Table of

We present a framework to understand GAN training as alternating density ratio estimation, and approximate divergence minimization. This provides an interpretation for the mismatched GAN generator and discriminator objectives often used in practice, and explains the problem of poor sample diversity. Further, we derive a family of generator objectives that target arbitrary f -divergences without...

In this paper, we propose Dynamics Transfer GAN; a new method for generating video sequences based on generative adversarial learning. The spatial constructs of a generated video sequence are acquired from the target image. The dynamics of the generated video sequence are imported from a source video sequence, with arbitrary motion, and imposed onto the target image. To preserve the spatial con...