Automatic essay grading (AEG) is a process in which machines assign a grade to an essay written in response to a topic, called the prompt. Zero-shot AEG is when we train a system to grade essays written to a new prompt which was not present in our training data. In this paper, we describe a solution to the problem of zero-shot automatic essay grading, using cognitive information, in the form of gaze behaviour. Our experiments show that using gaze behaviour helps in improving the performance of AEG systems, especially when we provide a new essay written in response to a new prompt for scoring, by an average of almost 5 percentage points of QWK.

A key challenge for reinforcement learning is solving long-horizon planning and control problems. Recent work has proposed leveraging programs to help guide the learning algorithm in these settings. However, these approaches impose a high manual burden on the user since they must provide a guiding program for every new task they seek to achieve. We propose an approach that leverages program synthesis to automatically generate the guiding program. A key challenge is how to handle partially observable environments. We propose model predictive program synthesis, which trains a generative model to predict the unobserved portions of the world, and then synthesizes a program based on samples from this model in a way that is robust to its uncertainty. We evaluate our approach on a set of challenging benchmarks, including a 2D Minecraft-inspired ``craft'' environment where the agent must perform a complex sequence of subtasks to achieve its goal, a box-world environment that requires abstract reasoning, and a variant of the craft environment where the agent is a MuJoCo Ant. Our approach significantly outperforms several baselines, and performs essentially as well as an oracle that is given an effective program.

Transformers are arguably the main workhorse in recent Natural Language Processing research. By definition a Transformer is invariant with respect to reorderings of the input. However, language is inherently sequential and word order is essential to the semantics and syntax of an utterance. In this paper, we provide an overview of common methods to incorporate position information into Transformer models. The objectives of this survey are to i) showcase that position information in Transformer is a vibrant and extensive research area; ii) enable the reader to compare existing methods by providing a unified notation and meaningful clustering; iii) indicate what characteristics of an application should be taken into account when selecting a position encoding; iv) provide stimuli for future research. The Transformer model as introduced by Vaswani et al. (2017) has been found to perform well for many tasks, such as machine translation or language modeling. With the rise of pretrained language models (PLMs) (Peters et al., 2018; Howard & Ruder, 2018; Devlin et al., 2019; Brown et al., 2020) Transformer models have become even more popular. As a result they are at the core of many state of the art natural language processing (NLP) models. A Transformer model consists of several layers, or blocks. Each layer is a self-attention (Vaswani et al., 2017) module followed by a feed-forward layer. Layer normalization and residual connections are additional components of a layer.

The Shapley value has become popular in the Explainable AI (XAI) literature, thanks, to a large extent, to a solid theoretical foundation, including four "favourable and fair" axioms for attribution in transferable utility games. The Shapley value is provably the only solution concept satisfying these axioms. In this paper, we introduce the Shapley value and draw attention to its recent uses as a feature selection tool. We call into question this use of the Shapley value, using simple, abstract "toy" counterexamples to illustrate that the axioms may work against the goals of feature selection. From this, we develop a number of insights that are then investigated in concrete simulation settings, with a variety of Shapley value formulations, including SHapley Additive exPlanations (SHAP) and Shapley Additive Global importancE (SAGE).

The goal of person search is to localize and match query persons from scene images. For high efficiency, one-step methods have been developed to jointly handle the pedestrian detection and identification sub-tasks using a single network. There are two major challenges in the current one-step approaches. One is the mutual interference between the optimization objectives of multiple sub-tasks. The other is the sub-optimal identification feature learning caused by small batch size when end-to-end training. To overcome these problems, we propose a decoupled and memory-reinforced network (DMRNet). Specifically, to reconcile the conflicts of multiple objectives, we simplify the standard tightly coupled pipelines and establish a deeply decoupled multi-task learning framework. Further, we build a memory-reinforced mechanism to boost the identification feature learning. By queuing the identification features of recently accessed instances into a memory bank, the mechanism augments the similarity pair construction for pairwise metric learning. For better encoding consistency of the stored features, a slow-moving average of the network is applied for extracting these features. In this way, the dual networks reinforce each other and converge to robust solution states. Experimentally, the proposed method obtains 93.2% and 46.9% mAP on CUHK-SYSU and PRW datasets, which exceeds all the existing one-step methods.

This paper reports on progress integrating the speech recognition toolkit ESPnet into Elpis, a web front-end originally designed to provide access to the Kaldi automatic speech recognition toolkit. The goal of this work is to make end-to-end speech recognition models available to language workers via a user-friendly graphical interface. Encouraging results are reported on (i) development of an ESPnet recipe for use in Elpis, with preliminary results on data sets previously used for training acoustic models with the Persephone toolkit along with a new data set that had not previously been used in speech recognition, and (ii) incorporating ESPnet into Elpis along with UI enhancements and a CUDA-supported Dockerfile.

This work presents a novel pipeline that demonstrates what is achievable with a combined effort of state-of-the-art approaches, surpassing the 50% exact match on NaturalQuestions and EfficentQA datasets. Specifically, it proposes the novel R2-D2 (Rank twice, reaD twice) pipeline composed of retriever, reranker, extractive reader, generative reader and a simple way to combine them. Furthermore, previous work often comes with a massive index of external documents that scales in the order of tens of GiB. This work presents a simple approach for pruning the contents of a massive index such that the open-domain QA system altogether with index, OS, and library components fits into 6GiB docker image while retaining only 8% of original index contents and losing only 3% EM accuracy.

Autism spectrum disorder is a developmental disorder characterized by significant social, communication, and behavioral challenges. Individuals diagnosed with autism, intellectual, and developmental disabilities (AUIDD) typically require long-term care and targeted treatment and teaching. Effective treatment of AUIDD relies on efficient and careful behavioral observations done by trained applied behavioral analysts (ABAs). However, this process overburdens ABAs by requiring the clinicians to collect and analyze data, identify the problem behaviors, conduct pattern analysis to categorize and predict categorical outcomes, hypothesize responsiveness to treatments, and detect the effects of treatment plans. Successful integration of digital technologies into clinical decision-making pipelines and the advancements in automated decision-making using Artificial Intelligence (AI) algorithms highlights the importance of augmenting teaching and treatments using novel algorithms and high-fidelity sensors. In this article, we present an AI-Augmented Learning and Applied Behavior Analytics (AI-ABA) platform to provide personalized treatment and learning plans to AUIDD individuals. By defining systematic experiments along with automated data collection and analysis, AI-ABA can promote self-regulative behavior using reinforcement-based augmented or virtual reality and other mobile platforms. Thus, AI-ABA could assist clinicians to focus on making precise data-driven decisions and increase the quality of individualized interventions for individuals with AUIDD.

(Maher 2012) introduced an approach for relative expressiveness of defeasible logics, and two notions of relative expressiveness were investigated. Using the first of these definitions of relative expressiveness, we show that all the defeasible logics in the DL framework are equally expressive under this formulation of relative expressiveness. The second formulation of relative expressiveness is stronger than the first. However, we show that logics incorporating individual defeat are equally expressive as the corresponding logics with team defeat. Thus the only differences in expressiveness of logics in DL arise from differences in how ambiguity is handled. This completes the study of relative expressiveness in DL begun in \cite{Maher12}.

This week's biggest story continues to be the Perseverance rover. NASA's latest space robot has brought another Linux device to Mars, and is already sending back some impressive pictures. We'll have to wait a little longer for HD video and the first drone flight -- beware of fake videos circulating on social media -- but next week should be even better. Until then you can always catch up on WandaVision's bite-size episodes, and make sure you stick around after the credits start to roll. Blizzard's online-only convention is going on this weekend, and the opening keynotes provided plenty of info about upcoming games.