Our work here is motivated by a problem faced by our lo- cal food-bank (Food Bank for the Southern Tier of New York (FBST)) in operating their mobile food pantry program. Every day, FBST uses a truck to deliver food supplies directly to distribution sites (soup kitchens/pantries/etc.). When the truck arrives at a site, the operator observes the demand there and chooses how much to allocate before moving to the next site. The number of people assembling at each site changes from day to day, and the operator typically does not know the demand of later sites (but has a sense of the demand distribution based on previous visits). Finally, the amount of food in the truck is usually insufficient to meet the total demand, and so the operator must under-allocate at each site, while trying to be fair across all sites. The question is: What is a fair allocation here, and how can it be computed? In offline problems, where demands (more generally, utility functions) for all agents are known to the principal, there are many well-studied notions of fair allocation of limited re- sources. A relevant notion in our context is that a fair allocation is one satisfying two desiderata: pareto-efficiency (for any agent to benefit, another must be hurt) and envy-freeness (no agent prefers an allocation received by another). This definition draws its importance from the fact that in many al- location settings, it is both known to be achievable, and also to encompass other natural desiderata (in particular, proportionality, wherein each agent’s utility is at least that achieved under equal allocation). In particular, when goods are divisible, then for a large class of utility functions, an allocation satisfying both is easily computed (via a convex optimization program) by maximizing the Nash Social Welfare (NSW) objective subject to allocation constraints. Many settings, much like the FBST operating their mo- bile food pantry, have principals make decisions online, with incomplete knowledge on the demands for agents to come. However, these principals have access to historical data al- lowing them to generate demand histograms for each agent. Designing allocation algorithms in this setting necessitates utilizing the Bayesian information of the demand distribution to ensure equitable access to the resource, while adapting to the online realization of demands as it unfolds. Guaranteeing pareto-efficiency and envy-freeness simultaneously is impossible in this setting. However, it is important to develop algorithms which achieve probabilistic version of fairness by utilizing the distributional knowledge to develop algorithms that are approximately fair.
In this paper, we present new results on the fair and efficient allocation of indivisible goods to agents whose preferences correspond to matroid rank functions. This is a versatile valuation class, with several desirable properties (monotonicity, submodularity) which naturally models several real-world domains. We use these properties to our advantage: first, we show that when agent valuations are matroid rank functions, a socially optimal (i.e. utilitarian social welfare-maximizing) allocation that achieves envy-freeness up to one item (EF1) exists and is computationally tractable. We also prove that the Nash welfare-maximizing and the leximin allocations both exhibit this fair- ness/efficiency combination, by showing that they can be achieved by minimizing any symmetric strictly convex function of agents’ valuations over utilitarian optimal outcomes. Moreover, for a subclass of these valuation functions based on maximum (unweighted) bipartite matching, we show that a leximin allocation can be computed in polynomial time.
The analysis of satellite imagery will prove a crucial tool in the pursuit of sustainable development. While Convolutional Neural Networks (CNNs) have made large gains in natural image analysis, their application to multi-spectral satellite images (wherein input images have a large number of channels) remains relatively unexplored. In this paper, we compare different methods of leveraging multi-band information with CNNs, demonstrating the performance of all compared methods on the task of semantic segmentation of agricultural vegetation (vineyards). We show that standard industry practice of using bands selected by a domain ex- pert leads to a significantly worse test accuracy than the other methods compared. Specifically, we com- pare: using bands specified by an expert; using all available bands; learning attention maps over the input bands; and leveraging Bayesian optimisation to dictate band choice. We show that simply using all available band information already increases test time performance, and show that the Bayesian optimisation, novelly applied to band selection in this work, can be used to further boost accuracy.
A mechanism for predicting whether individual regions will meet there UN Sustainability for Development Goals (SDGs) is presented which takes into consideration the potential relationships be- tween time series associated with individual SDGs, unlike previous work where an independence assumption was made. The challenge is in identifying the relationships and then using these relationships to make SDG attainment predictions. To this end, the SDG Relational Multivariate Forecast- ing (SDG-RMF) attainment prediction methodology is presented. A multivariate forecasting mechanism for forecasting SDGs time series The results demonstrate that by considering the relationships between time series, more accurate SDG forecast predictions can be made.
While artificial intelligence (AI) has been heralded as a technology capable of solving unique problems, social good challenges are inherently structural and require the partnership of many stake- holders in order to apply AI for social good (AI4SG) in a sustainable and scaled manner. This paper explains current challenges in project implementation, surveys framework approaches, and contributes our differentiating lessons learned on scaling projects to problem domain-wide impact. The goal is to guide partnering organizations through challenges and identifying opportunities to accelerate the application of AI4SG.
The alarming trend of increasing mental health problems and the global inability to find effective ways to address them is hampering both individual and societal good. Barriers to access mental health care are many and high, ranging from socio- economic inequalities to personal stigmas. This gives technology, especially technology based in artificial intelligence, the opportunity to help alleviate the situation and offer unique possibilities to tackle the problem. The multi- and interdisciplinary research on persuasive technology, which attempts to change behavior or attitudes without deception or coercion, shows promise in improving wellbeing, which results in increased equality and social good. This paper presents such systems with a brief overview of the field, and offers general, technical and critical thoughts on the implementation as well as impact. We believe that such technology can complement existing mental health care solutions to reduce inequalities in access as well as inequalities resulting from the lack of it.
Fairness, in machine learning research, is often conceived as an exercise in constrained optimization, based on a predefined fairness metric. We argue that this abstract model of algorithmic fairness is a poor match for the real-world, in which applications are likely to be embedded within a larger context involving multiple classes of stakeholders as well as multiple social and technical systems. We may expect multiple, competing claims around fairness coming from various stakeholders, especially in applications oriented towards social good. We propose that computational social choice is a promising framework for the integration of multiple perspectives on system outcomes in fairness- aware systems and provide an example case of personalized recommendation for a non-profit.
As fairness and discrimination concerns permeate the design of both machine learning algorithms and mechanism design problems, we discuss differences in approaches between these two fields. We aim to bridge these two communities into a cohesive narrative that en- compasses both the large-scale capabilities of machine learning and group-focused fairness as well as the strategic incentives and utility- based notions of fairness from mechanism de- sign, showing their necessity in designing a fair pipeline.
The COVID-19 pandemic has been a major challenge to humanity, with 12.7 million confirmed cases as of July 13th, 2020 . In previous work, we described how Artificial Intelligence can be used to tackle the pandemic with applications at the molecular, clinical, and societal scales . In the present follow-up article, we review these three research directions, and assess the level of maturity and feasibility of the approaches used, as well as their potential for operationalization. We also summarize some commonly encountered risks and practical pitfalls, as well as guidelines and best practices for formulating and deploying AI applications at different scales.
Word embeddings learned from collections of data have demonstrated a significant level of biases. When these embeddings are used in machine learn- ing tasks it often amplifies the bias. We propose a debiasing method that uses (Figure 1) a hybrid classification - variational autoencoder network. In this work, we developed a semi-supervised classification algorithm based on variational autoencoders which learns the latent structure within the dataset and then based on learned latent structure adaptively re-weights the importance of certain data points while training. Experimental results have shown that the proposed approach works better than existing SoTA methods for debiasing word embeddings.
Not all patients who need kidney transplant can find a donor with compatible characteristics. Kidney exchange programs (KEPs) seek to match such incompatible patient-donor pairs together, usually with the objective of maximizing the total number of transplants. We propose a randomized policy for selecting an optimal solution in which patients’ equity of opportunity to receive a transplant is promoted. Our approach gives rise to the problem of enumerating all optimal solutions, which we tackle using a hybrid of constraint programming and linear programming. We empirically demonstrate the advantages of our proposed method over the common practice of using the first optimal solution obtained by a solver.
The increasingly widespread use of Natural Language Processing (NLP) in AI applications must be continually monitored for biases and false associations, especially those surrounding protected or disadvantaged classes of people. We discuss methods and algorithms used to mitigate such biases and their weak points, using real world examples in civilian agencies of the US government.
Deep Learning is changing the face of agriculture. Combined with high-resolution aerial imagery, these methods enable farmers to understand and manage their farms with previously unseen precision and efficiency. Beyond reducing costs for an industry already under significant economic stress, these advances have key environmental benefits as well: maximizing production, reducing waste, anticipating disruptions to supply chains, and limiting the use of chemicals and water through targeted application. Our approach uses a U-net based neural network to predict the density of flowering pineapple plants from aerial imagery, enabling farmers to optimize their harvesting schedule.
In this paper we consider the problem of forming student teams adequate for company internship tasks. First, we provide a formalisation of the Feasible Team-For-Task Allocation Problem, and show the computational hardness of solving it optimally. Thereafter, we propose Edu2Com, an any- time heuristic algorithm that generates an initial team allocation that is then improved in an iterative process. Finally, we conduct a systematic evaluation and show that Edu2Com manages to (a) out- perform CPLEX in computation time, and (b) reach optimality, in the experiments considered.
Artificial Intelligence (AI) has started to transform our economy and society, more specifically, AI has the potential to make both labor and machines more productive while displacing certain human tasks and simultaneously introducing new tasks into the economy. Using online job postings data, this pa- per proposes novel methodologies to characterize the dynamic evolution of occupational task-share demands across different industries in the U.S. la- bor market and estimates the implied US-$ market values for skills. The paper develops a multi-variate and multi-step long short term memory (LSTM) network architecture to estimate 12-month and 24- month ahead forecasts of task-shares with 10% root mean-squared error. The industry-specific insights on occupation evolution and forecasts on task- shares will facilitate the policy-makers and strategy leaders’ decision-making to transform the current workforce for the future.
Staying up-to-date with current medical research can be a challenge for doctors and other medi- cal decision-makers. Systematic reviews are one of the key tools that doctors use to stay informed. These are meta-analyses of all the relevant research with the intention of answering one specific ques- tion within the healthcare domain. Cochrane pro- duces systematic reviews of medical research that are globally considered as a gold standard for high- quality healthcare information. However, because of the high volume of papers published and the fact that Cochrane’s review authors are volunteers, it can take up to three years to write and publish one of these reviews. Our research focuses on speeding up this process. We propose a hybrid human-AI system to establish the topical area of a newly pub- lished paper faster, easing the process of searching for papers to include in a review.
Conflict resolution practitioners consistently struggle with access to structured armed conflict data, a dataset already rife with uncertainty, inconsistency, and politicization. Due to the lack of a standardized approach to collating conflict data, publicly available armed conflict datasets often require manipulation depending upon the needs of end users. Transformation of armed conflict data tends to be a manual, time consuming task that nonprofits with limited budgets struggle to keep up with. In this paper, we explore the use of a deep natural language processing (NLP) model to aid the transformation of armed conflict data for conflict analysis. Our model drastically reduces the time spent on manual data transformations and improves armed conflict event classification by identifying multiple incidence types. This minimizes the human supervision cost and allows nonprofits to access a broader range of conflict data sources to reduce reporting bias. Thus our model contributes to the incorporation of technology in the peace building and conflict resolution sector.
Detection of burn marks due to wildfires in inaccessible rain forests is important for various disaster management and ecological studies. The fragmented nature of arable landscapes and diverse
cropping patterns often thwart the precise mapping
of burn scars. Recent advances in remote-sensing
and availability of multimodal data offer a viable
solution to this mapping problem. However, the
task to segment burn marks is difficult because of
its indistinguishably with similar looking land patterns, severe fragmented nature of burn marks and
partially labelled noisy datasets.
In this work we present AmazonNET – a convolutional based network that allows extracting of burn
patters from multimodal remote sensing images.
The network consists of UNet- a well-known encoder decoder type of architecture with skip connections. The proposed framework utilises stacked
RGB-NIR channels to segment burn scars from the
pastures by training on a new weakly labelled noisy
dataset from Amazonia.
Our model illustrates superior performance by correctly identifying partially labelled burn scars and
rejecting incorrectly labelled samples, demonstrating our approach as one of the first to effectively
utilise deep learning based segmentation models in
multimodal burn scar identification.
Within the relatively busy area of fair machine learning that has been dominated by classification fairness research, fairness in clustering has started to see some recent attention. In this position paper, we assess the existing work in fair clustering and observe that there are several directions that are yet to be explored, and postulate that the state-of-the- art in fair clustering has been quite parochial in out- look. We posit that widening the normative prin- ciples to target for, characterizing shortfalls where the target cannot be achieved fully, and making use of knowledge of downstream processes can significantly widen the scope of research in fair clustering research. At a time when clustering and unsupervised learning are being increasingly used to make and influence decisions that matter significantly to human lives, we believe that widening the ambit of fair clustering is of immense significance.
As disasters such as earthquakes and floods be- come more frequent and detrimental, it is increas- ingly important that water infrastructure resilience be strategically enhanced to support post-disaster functionality and recovery. In this paper, we focus on the problem of strategically building seismic- resilient pipe networks to ensure direct water sup- ply to critical customers and certain proximity to water sources for residential areas, which we for- malize as the Steiner network problem with cov- erage constraints. We present an efficient mixed- integer linear program encoding to solve the prob- lem. We also investigate the problem of planning partial network installments to maximize efficiency over time and propose an effective sequential plan- ning algorithm to solve it. We evaluate our algo- rithms on synthetic water networks and apply them to a case study on a water service zone in Los An- geles, which demonstrate the effectiveness of our methods for large-scale real-world applications.