For many organisations and their workers, technology automation is an opportunity and a threat. On one hand, using robotic process automation and other tools creates efficiencies by automating manual and repetitive tasks. On the other, it has the potential to make jobs redundant. Tech execs gathered at roundtable events in Sydney and Melbourne recently to discuss the impact automation technologies are having on their organisations. A key issue raised during the discussion was the challenge of ensuring governance policies and procedures are followed as'bots' replace tasks carried out by humans across the business.

Many currently deployed Reinforcement Learning agents work in an environment shared with humans, be them co-workers, users or clients. It is desirable that these agents adjust to people's preferences, learn faster thanks to their help, and act safely around them. We argue that most current approaches that learn from human feedback are unsafe: rewarding or punishing the agent a-posteriori cannot immediately prevent it from wrong-doing. In this paper, we extend Policy Gradient to make it robust to external directives, that would otherwise break the fundamentally on-policy nature of Policy Gradient. Our technique, Directed Policy Gradient (DPG), allows a teacher or backup policy to override the agent before it acts undesirably, while allowing the agent to leverage human advice or directives to learn faster. Our experiments demonstrate that DPG makes the agent learn much faster than reward-based approaches, while requiring an order of magnitude less advice.

Decentralized machine learning is a promising emerging paradigm in view of global challenges of data ownership and privacy. We consider learning of linear classification and regression models, in the setting where the training data is decentralized over many user devices, and the learning algorithm must run on-device, on an arbitrary communication network, without a central coordinator. We propose COLA, a new decentralized training algorithm with strong theoretical guarantees and superior practical performance. Our framework overcomes many limitations of existing methods, and achieves communication efficiency, scalability, elasticity as well as resilience to changes in data and participating devices.

In mammals this incorporates the cortex, the hippocampus, the claustrum, the amygdala, the basal ganglia, and the olfactory bulb. Convergent evolution results in analogous characters with similar appearances or functions, although these were not present in the last common ancestor of the two lineages. Most species are characterized by a high brain-to-body mass ratio, ecological flexibility, and a complex social life, featuring long-term partnerships and dynamic groups structured by social relationships. The term derives from the Greek word hodos which means'road'. Each layer is constituted by distinctive cell populations with unique connectivity patterns. At first glance, neocortical lamination looks uniform (and is therefore sometimes called'isocortical').

Putting aside the concept of'personality' and its disputed nature within psychology, the new platform attempts to group subjects into different personality types on the basis of eye-motion. The research comes from the University of South Australia. Visual exploration is driven by two main factors. First there is the stimuli in our environment; and second, in response to our own individual interests and intentions. It is through the latter that some researchers think that personality traits can be discerned.

Keeping a cool head and discerning what's beneficial and what is not are indispensable considerations to avoid a fiasco Today the role of the CIO is perceived as a strategic business position, tasked with driving change and transformation within their organisation instead of exclusively focusing on IT operations. It's one of the most dynamic executive roles which requires constant development and adaptation to digital ecosystems, business model innovations, technology demands and stakeholders' expectations. Adapting to tight budgets and focusing on internal operations are now just part of a long list of duties expected from a CIO. Driving business operations and leading innovation and transformation are regarded as extremely desired skills. In an interview with our sister title CIO UK, SThree CIO Lance Fisher explained the change in nature of his job: "I split IT into two. You have to learn and understand the business and its competitive advantage, understand how your competitors run their businesses and then look to deliver competitive advantage. I firmly believe as a CIO you are the glue. You need to connect IT to the business, vendors and your clients."

Deep learning models have achieved remarkable success in natural language inference (NLI) tasks. While these models are widely explored, they are hard to interpret and it is often unclear how and why they actually work. In this paper, we take a step toward explaining such deep learning based models through a case study on a popular neural model for NLI. In particular, we propose to interpret the intermediate layers of NLI models by visualizing the saliency of attention and LSTM gating signals. We present several examples for which our methods are able to reveal interesting insights and identify the critical information contributing to the model decisions.

Computational modeling of human multimodal language is an emerging research area in natural language processing spanning the language, visual and acoustic modalities. Comprehending multimodal language requires modeling not only the interactions within each modality (intra-modal interactions) but more importantly the interactions between modalities (cross-modal interactions). In this paper, we propose the Recurrent Multistage Fusion Network (RMFN) which decomposes the fusion problem into multiple stages, each of them focused on a subset of multimodal signals for specialized, effective fusion. Cross-modal interactions are modeled using this multistage fusion approach which builds upon intermediate representations of previous stages. Temporal and intra-modal interactions are modeled by integrating our proposed fusion approach with a system of recurrent neural networks. The RMFN displays state-of-the-art performance in modeling human multimodal language across three public datasets relating to multimodal sentiment analysis, emotion recognition, and speaker traits recognition. We provide visualizations to show that each stage of fusion focuses on a different subset of multimodal signals, learning increasingly discriminative multimodal representations.

Professor Phil Cohen, Director of the Laboratory for Dialogue Research, at Monash University, Faculty of Information Technology, will be a keynote speaker at the Digital A.I Summit, in Melbourne on August 28, 2018. He will discuss the strengths and weaknesses of current "conversational A.I" technology, which is used in today's digital assistants (e.g., Apple Computer's Siri, Google's Google Now, IBM's Watson, and Microsoft's Cortana, etc.), as well as for general classes of applications such as customer service. He will provide a way forward that promises to overcome their limitations to build systems that can participate in flexible, collaborative dialogues that help users to achieve their goals. The Digital A.I Summit is Australia's premier event for bringing together industry leaders and provides a collection of exclusive keynotes and expert panels explaining AI and how it is rapidly transforming many Australian workplaces to significantly improve their productivity. The Digital A.I Summit is a one-day conference and is run in conjunction with the Victorian Government's Digital Innovation Festival.

In this technical report, we present jLDADMM---an easy-to-use Java toolkit for conventional topic models. jLDADMM is released to provide alternatives for topic modeling on normal or short texts. It provides implementations of the Latent Dirichlet Allocation topic model and the one-topic-per-document Dirichlet Multinomial Mixture model (i.e. mixture of unigrams), using collapsed Gibbs sampling. In addition, jLDADMM supplies a document clustering evaluation to compare topic models. jLDADMM is open-source and available to download at: https://github.com/datquocnguyen/jLDADMM