Spain

ECAI 2020

FishNet: A Unified Embedding for Salmon Recognition

block diagonal structure

network compression

neural network

ECAI2020 featured a full programme of technical papers, as well as many AI-related events of interest for researchers, students and all attendants who are interested on contemporary AI, including:




**An extensive programme of workshops, tutorials, and invited speakers**

**The Prestigious Applications of Intelligent Systems conference (PAIS)**

**The Frontiers in AI track sessions**

**The Lab To Market event**

**The Women in AI time**

**The EU Challenges forum**

… and many more!

ECAI 2020 featured a full programme of technical papers, as well as many AI-related events of interest for researchers, students and all attendants who are interested on contemporary AI.

technical paper

In this work we consider optimization problems that require to make interdependent offline and online decisions under uncertainty. We broadly refer to long-term strategic decisions as offline and to short-term operational decisions as online. For example, in Distributed Energy Management Systems we may need to define (offline) a daily production schedule for an industrial plant, and then manage (online) its power supply on a hour by hour basis. Traditionally offline and online phases are tackled in isolation, leading to some drawbacks: offline decisions are taken without regard for the capabilities of the downstream online solver; while the applicability of the best approaches for online decisions (e.g. anticipatory algorithms) is limited by the need to provide high responsiveness. Starting from a (literature-based) baseline, we define general methods for leading to significant quality improvements, at the cost of an increased computation effort either in the offline or the online phase. All our methods have broad applicability, and provide multiple options to balance the solution quality/time trade-off, suiting a variety of practical application scenarios with both offline and online decisions and featuring continuous and discrete decisions. An extensive analysis of the experimental results shows that offline/online integration may lead to substantial benefits.

Hybrid Offline/Online Optimization Under Uncertainty

In the field of Information Retrieval, word embedding models have shown to be effective in several tasks. In this paper, we show how one of these neural embedding techniques can be adapted to the recommendation task. This adaptation only makes use of collaborative filtering information, and the results show that it is able to produce effective recommendations efficiently.

Shallow Neural Models for Top-N Recommendation

Sketches have been employed since the ancient era of cave paintings for simple illustrations to represent real-world entities. The abstract nature and varied artistic styling make automatic recognition of drawings more challenging than other areas of image classification. Moreover, the representation of sketches as a sequence of strokes instead of raster images introduces them at the correct abstract level. However, dealing with images as a sequence of small information makes it challenging. In this paper, we propose a Transformer-based network, dubbed as TransSketchNet, for sketch recognition. This architecture incorporates ordinal information to perform the classification task in real-time through vector images.

TransSketchNet: Attention-based Sketch Recognition using Transformers

The next technological revolution will be interwoven to the proliferation of intelligent systems. As we bridge the gap between physical and cyber worlds, we will give rise to large-scale, multi-agent based technologies. A key challenge that cities of the future will have to face is coordination in the use of limited resources, central to which is finding an optimal allocation between agents. To truly allow for scalable solutions, we need to shift from traditional approaches, to multi-agent solutions, ideally run on-device. We highlight a novel heuristic (ALMA), which exhibits such properties, for solving the assignment problem. ALMA is decentralized, requires only partial feedback, and has constant in the total problem size running time, under reasonable assumptions on the preference domain of the agents, making it ideal for an on-device implementation. We have evaluated ALMA in a variety of scenarios including synthetic and real data, time constraints, and on-line settings. In all of the cases, ALMA was able to reach high social welfare, while being orders of magnitude faster than the centralized, optimal algorithm.

Efficient Allocations in Constant Time: Towards Scalable Solutions in the Era of Large Scale Intelligent Systems

This paper highlights a new learning model that introduces a contextual dimension to the well-known Q-Learning algorithm. Through the identification of different contexts, the learning process is adapted accordingly, thus converging to enhanced results. The proposed learning model includes a simulated annealing (SA) process that accelerates the convergence process. The model is integrated in a multi-agent decision support system for electricity market players negotiations, enabling the experimentation of results using real electricity market data

Contextual Q-Learning

"How can Artificial Intelligence help a stateless minority from online abuse? Research efforts in hate speech detection thus far have largely focused on identifying and subsequently filtering out negative content that specifically targets them. In this paper, we highlight a recent work that tackles a different aspect of web-vulnerability of marginalized communities: sparsity of pro-minority voices championing their cause. The highlighted paper advocates that blocking hate alone may not be sufficient in these cases as the internet shapes community perception to a great extent in modern times and supportive comments to a vulnerable community serve
 a different purpose. Using an Active Sampling approach, the paper constructs a nuanced voice-for-the-voiceless classifier that automatically discovers comments supporting an (allegedly) persecuted minority. In the context of the Rohingya refugee crisis, one of the biggest humanitarian crises of modern times, the paper presents promising results that can substantially aid content moderation efforts in finding positive content supporting the Rohingyas."

The Refugee Experience Online: Surfacing Positivity Amidst Hate

Modelling a problem as finding a shortest path in a graph has been successful in many AI problems. However, the classical shortest path framework might not be suitable for domains such as planning a career pathway optimized to the skills needed along the way. To address this gap, we introduce graphs with item sets, where each node has a set of items from a finite, global set. We define a cost function for paths in graphs with item sets and show that it lacks the prefix optimality property. We further define two decision problems inspired from career planning. We show that these are NP-complete, unlike the traditional shortest path problem, which is solved in polynomial time. Our study gives a theoretical understanding that could be used as a basis for designing efficient solving approaches to the problems studied.

The Challenge of Optimal Paths in Graphs with Item Sets

As "Big Data" has become pervasive, an increasing amount of research has connected the dots between human behaviour in the offline and online worlds. Consequently, researchers have exploited these new findings to create models that better predict different aspects of human life and recommend future behaviour. To date, however, we do not yet fully understand the similarities and differences of human behaviour in these virtual and physical worlds. Here, we analyse and discuss the mobility and application usage of 400,000 individuals over eight months. We find an astonishing similarity between people's mobility in the physical space and how they move from app to app in smartphones. Our data shows that individuals use and visit a finite number of apps and places, but they keep exploring over time. In particular, two distinct profiles of individuals emerge: those that keep changing places and services, and those that are stable over time, named as "explorers" and "keepers". We see these findings as crucial to enrich a discussion for the potentials and the challenges of building human-centric AI systems, which might leverage recent results in Computational Social Science.

Understanding individual behaviour: from virtual to physical patterns

The prevailing methodology for integrating decision making and ethics is to modify autonomous systems in an ad hoc way to incorporate moral sensibility. However, these provisional modifications often lead to behavior that jeopardizes the intentions of developers or the values of stakeholders. We propose a novel approach for building moral autonomous systems that optimally completes a task and follows an ethical framework by decoupling ethical compliance from task completion. This paper offers a formal definition of our approach along with its key properties, an example based on prima facie duties, and a demonstration that uses our open source library.

An Integrated Approach to Moral Autonomous Systems

Contracts are the cornerstone of legal agreements in the industry. As essential legal documents, contracts are subject to negotiations, demanding extensive analysis and evaluation efforts. Although the emergence of machine learning has enabled assistance tools for text analysis tasks, the specificity and constraints of each
 business context remain obstacles for the automation of contracts evaluation. In this paper, we propose an AutoML based approach for automated negotiation assistance that uses expert annotated contracts and the business-specific knowledge of acceptance policies. Driven by policies rules, our approach generates a classification process composed of hierarchies of complementary classifiers, each being automatically prepared according, but not limited to, feature extraction, learning model and data granularity. Experiments conducted on real-world service contracts have yielded promising results.

Downloads

Next from ECAI 2020

Hybrid Offline/Online Optimization Under Uncertainty

Similar lecture

On Minimizing Diagonal Block-wise Differences for Neural Network Compression

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from ECAI 2020

Hybrid Offline/Online Optimization Under Uncertainty

Similar lecture

On Minimizing Diagonal Block-wise Differences for Neural Network Compression

Downloads