Hopfield networks are artificial neural networks which store memory patterns on the states of their neurons by choosing recurrent connection weights and update rules such that the energy landscape of the network forms attractors around the memories. How many stable, sufficiently-attracting memory patterns can we store in such a network using $N$ neurons? The answer depends on the choice of weights and update rule. Inspired by setwise connectivity in biology, we extend Hopfield networks by adding setwise connections and embedding these connections in a simplicial complex. Simplicial complexes are higher dimensional analogues of graphs which naturally represent collections of pairwise and setwise relationships. We show that our simplicial Hopfield networks increase memory storage capacity. Surprisingly, even when connections are limited to a small random subset of equivalent size to an all-pairwise network, our networks still outperform their pairwise counterparts. Such scenarios include non-trivial simplicial topology. We also test analogous modern continuous Hopfield networks, offering a potentially promising avenue for improving the attention mechanism in Transformer models.

Over the last years, significant advances have been made in robotic manipulation, but still, the handling of non-rigid objects, such as cloth garments, is an open problem. Physical interaction with non-rigid objects is uncertain and complex to model. Thus, extracting useful information from sample data can considerably improve modeling performance. However, the training of such models is a challenging task due to the high-dimensionality of the state representation. In this paper, we propose Controlled Gaussian Process Dynamical Model (CGPDM) for learning high-dimensional, nonlinear dynamics by embedding it in a low-dimensional manifold. A CGPDM is constituted by a low-dimensional latent space, with an associated dynamics where external control variables can act and a mapping to the observation space. The parameters of both maps are marginalized out by considering Gaussian Process (GP) priors. Hence, a CGPDM projects a high-dimensional state space into a smaller dimension latent space, in which it is feasible to learn the system dynamics from training data. The modeling capacity of CGPDM has been tested in both a simulated and a real scenario, where it proved to be capable of generalizing over a wide range of movements and confidently predicting the cloth motions obtained by previously unseen sequences of control actions.

In modern machine learning, inner product attention computation is a fundamental task for training large language models such as Transformer, GPT-1, BERT, GPT-2, GPT-3 and ChatGPT. Formally, in this problem, one is given as input three matrices $Q, K, V \in [-B,B]^{n \times d}$, and the goal is to construct the matrix $\mathrm{Att}(Q,K,V) := \mathrm{diag}(A {\bf 1}_n)^{-1} A V \in \mathbb{R}^{n \times d}$, where $A = \exp(QK^\top/d)$ is the `attention matrix', and $\exp$ is applied entry-wise. Straightforward methods for this problem explicitly compute the $n \times n$ attention matrix $A$, and hence require time $\Omega(n^2)$ even when $d = n^{o(1)}$ is small. In this paper, we investigate whether faster algorithms are possible by implicitly making use of the matrix $A$. We present two results, showing that there is a sharp transition at $B = \Theta(\sqrt{\log n})$. $\bullet$ If $d = O(\log n)$ and $B = o(\sqrt{\log n})$, there is an $n^{1+o(1)}$ time algorithm to approximate $\mathrm{Att}(Q,K,V)$ up to $1/\mathrm{poly}(n)$ additive error. $\bullet$ If $d = O(\log n)$ and $B = \Theta (\sqrt{\log n})$, assuming the Strong Exponential Time Hypothesis from fine-grained complexity theory, it is impossible to approximate $\mathrm{Att}(Q,K,V)$ up to $1/\mathrm{poly}(n)$ additive error in truly subquadratic time $n^{2 - \Omega(1)}$. This gives a theoretical explanation for the phenomenon observed in practice that attention computation is much more efficient when the input matrices have smaller entries.

The objective of this work is to optimize the performance of a constant flow parallel mechanical displacement micropump, which has parallel pump chambers and incorporates passive check valves. The critical task is to minimize the pressure pulse caused by regurgitation, which negatively impacts the constant flow rate, during the reciprocating motion when the left and right pumps interchange their role of aspiration and transfusion. Previous works attempt to solve this issue via the mechanical design of passive check valves. In this work, the novel concept of overlap time is proposed, and the issue is solved from the aspect of control theory by implementing a RBF neural network trained by both unsupervised and supervised learning. The experimental results indicate that the pressure pulse is optimized in the range of 0.15 - 0.25 MPa, which is a significant improvement compared to the maximum pump working pressure of 40 MPa.

Machine learning models exhibit two seemingly contradictory phenomena: training data memorization, and various forms of forgetting. In memorization, models overfit specific training examples and become susceptible to privacy attacks. In forgetting, examples which appeared early in training are forgotten by the end. In this work, we connect these phenomena. We propose a technique to measure to what extent models "forget" the specifics of training examples, becoming less susceptible to privacy attacks on examples they have not seen recently. We show that, while non-convex models can memorize data forever in the worst-case, standard image, speech, and language models empirically do forget examples over time. We identify nondeterminism as a potential explanation, showing that deterministically trained models do not forget. Our results suggest that examples seen early when training with extremely large datasets - for instance those examples used to pre-train a model - may observe privacy benefits at the expense of examples seen later.

Span-based models are one of the most straightforward methods for named entity recognition (NER). Existing span-based NER systems shallowly aggregate the token representations to span representations. However, this typically results in significant ineffectiveness for long-span entities, a coupling between the representations of overlapping spans, and ultimately a performance degradation. In this study, we propose DSpERT (Deep Span Encoder Representations from Transformers), which comprises a standard Transformer and a span Transformer. The latter uses low-layered span representations as queries, and aggregates the token representations as keys and values, layer by layer from bottom to top. Thus, DSpERT produces span representations of deep semantics. With weight initialization from pretrained language models, DSpERT achieves performance higher than or competitive with recent state-of-the-art systems on eight NER benchmarks. Experimental results verify the importance of the depth for span representations, and show that DSpERT performs particularly well on long-span entities and nested structures. Further, the deep span representations are well structured and easily separable in the feature space.

Phone-level pronunciation scoring is a challenging task, with performance far from that of human annotators. Standard systems generate a score for each phone in a phrase using models trained for automatic speech recognition (ASR) with native data only. Better performance has been shown when using systems that are trained specifically for the task using non-native data. Yet, such systems face the challenge that datasets labelled for this task are scarce and usually small. In this paper, we present a transfer learning-based approach that leverages a model trained for ASR, adapting it for the task of pronunciation scoring. We analyze the effect of several design choices and compare the performance with a state-of-the-art goodness of pronunciation (GOP) system. Our final system is 20% better than the GOP system on EpaDB, a database for pronunciation scoring research, for a cost function that prioritizes low rates of unnecessary corrections.

Social media platforms are currently the main channel for political messaging, allowing politicians to target specific demographics and adapt based on their reactions. However, making this communication transparent is challenging, as the messaging is tightly coupled with its intended audience and often echoed by multiple stakeholders interested in advancing specific policies. Our goal in this paper is to take a first step towards understanding these highly decentralized settings. We propose a weakly supervised approach to identify the stance and issue of political ads on Facebook and analyze how political campaigns use some kind of demographic targeting by location, gender, or age. Furthermore, we analyze the temporal dynamics of the political ads on election polls.

It's something that many people are self-conscious of, and if you not a fan of your nose, we finally know who to blame. Scientists have revealed that Neanderthal DNA helps dictate the shape of your nose. A new study led by UCL researchers found that a particular gene, which leads to a taller nose, may have been the product of natural selection as ancient humans adapted to colder climates after leaving Africa. Dr Kaustubh Adhikari, who led the study, said: 'In the last 15 years, since the Neanderthal genome has been sequenced, we have been able to learn that our own ancestors apparently interbred with Neanderthals, leaving us with little bits of their DNA. 'Here, we find that some DNA inherited from Neanderthals influences the shape of our faces.

Transformers have been essential to pretraining success in NLP. While other architectures have been used, downstream accuracy is either significantly worse, or requires attention layers to match standard benchmarks such as GLUE. This work explores pretraining without attention by using recent advances in sequence routing based on state-space models (SSMs). Our proposed model, Bidirectional Gated SSM (BiGS), combines SSM layers with a multiplicative gating architecture that has been effective in simplified sequence modeling architectures. The model learns static layers that do not consider pair-wise interactions. Even so, BiGS is able to match BERT pretraining accuracy on GLUE and can be extended to long-form pretraining of 4096 tokens without approximation. Analysis shows that while the models have similar average accuracy, the approach has different inductive biases than BERT in terms of interactions and syntactic representations. All models from this work are available at https://github.com/jxiw/BiGS.