Search for: All records

Rankings are the primary interface through which many online platforms match users to items (e.g. news, products, music, video). In these two-sided markets, not only do the users draw utility from the rankings, but the rankings also determine the utility (e.g. exposure, revenue) for the item providers (e.g. publishers, sellers, artists, studios). It has already been noted that myopically optimizing utility to the users -- as done by virtually all learning-to-rank algorithms -- can be unfair to the item providers. We, therefore, present a learning-to-rank approach for explicitly enforcing merit-based fairness guarantees to groups of items (e.g. articles by the same publisher, tracks by the same artist). In particular, we propose a learning algorithm that ensures notions of amortized group fairness, while simultaneously learning the ranking function from implicit feedback data. The algorithm takes the form of a controller that integrates unbiased estimators for both fairness and utility, dynamically adapting both as more data becomes available. In addition to its rigorous theoretical foundation and convergence guarantees, we find empirically that the algorithm is highly practical and robust. 

Rankings are the primary interface through which many online platforms match users to items (e.g. news, products, music, video). In these two-sided markets, not only the users draw utility from the rankings, but the rankings also determine the utility (e.g. exposure, revenue) for the item providers (e.g. publishers, sellers, artists, studios). It has already been noted that myopically optimizing utility to the users – as done by virtually all learning-to-rank algorithms – can be unfair to the item providers. We, therefore, present a learning-to-rank approach for explicitly enforcing merit-based fairness guarantees to groups of items (e.g. articles by the same publisher, tracks by the same artist). In particular, we propose a learning algorithm that ensures notions of amortized group fairness, while simultaneously learning the ranking function from implicit feedback data. The algorithm takes the form of a controller that integrates unbiased estimators for both fairness and utility, dynamically adapting both as more data becomes available. In addition to its rigorous theoretical foundation and convergence guarantees, we find empirically that the algorithm is highly practical and robust. 

Iron ion batteries using Fe²⁺as a charge carrier have yet to be widely explored, and they lack high‐performing Fe²⁺hosting cathode materials to couple with the iron metal anode. Here, it is demonstrated that VOPO₄∙2H₂O can reversibly host Fe²⁺with a high specific capacity of 100 mAh g⁻¹and stable cycling performance, where 68% of the initial capacity is retained over 800 cycles. In sharp contrast, VOPO₄∙2H₂O's capacity of hosting Zn²⁺fades precipitously over tens of cycles. VOPO₄∙2H₂O stores Fe²⁺with a unique mechanism, where upon contacting the electrolyte by the VOPO₄∙2H₂O electrode, Fe²⁺ions from the electrolyte get oxidized to Fe³⁺ions that are inserted and trapped in the VOPO₄∙2H₂O structure in an electroless redox reaction. The trapped Fe³⁺ions, thus, bolt the layered structure of VOPO₄∙2H₂O, which prevents it from dissolution into the electrolyte during (de)insertion of Fe²⁺. The findings offer a new strategy to use a redox‐active ion charge carrier to stabilize the layered electrode materials.

 

Plating battery electrodes typically deliver higher specific capacity values than insertion or conversion electrodes because the ion charge carriers represent the sole electrode active mass, and a host electrode is unnecessary. However, reversible plating electrodes are rare for electronically insulating nonmetals. Now, a highly reversible iodine plating cathode is presented that operates on the redox couples of I₂/[ZnI_x(OH₂)_4−x]^2−xin a water‐in‐salt electrolyte. The iodine plating cathode with the theoretical capacity of 211 mAh g⁻¹plates on carbon fiber paper as the current collector, delivering a large areal capacity of 4 mAh cm⁻². Tunable femtosecond stimulated Raman spectroscopy coupled with DFT calculations elucidate a series of [ZnI_x(OH₂)_4−x]^2−xsuperhalide ions serving as iodide vehicles in the electrolyte, which eliminates most free iodide ions, thus preventing the consequent dissolution of the cathode‐plated iodine as triiodides.

 

Plating battery electrodes typically deliver higher specific capacity values than insertion or conversion electrodes because the ion charge carriers represent the sole electrode active mass, and a host electrode is unnecessary. However, reversible plating electrodes are rare for electronically insulating nonmetals. Now, a highly reversible iodine plating cathode is presented that operates on the redox couples of I₂/[ZnI_x(OH₂)_4−x]^2−xin a water‐in‐salt electrolyte. The iodine plating cathode with the theoretical capacity of 211 mAh g⁻¹plates on carbon fiber paper as the current collector, delivering a large areal capacity of 4 mAh cm⁻². Tunable femtosecond stimulated Raman spectroscopy coupled with DFT calculations elucidate a series of [ZnI_x(OH₂)_4−x]^2−xsuperhalide ions serving as iodide vehicles in the electrolyte, which eliminates most free iodide ions, thus preventing the consequent dissolution of the cathode‐plated iodine as triiodides.

 

Aqueous rechargeable batteries are promising solutions for large‐scale energy storage. Such batteries have the merit of low cost, innate safety, and environmental friendliness. To date, most known aqueous ion batteries employ metal cation charge carriers. Here, we report the first “rocking‐chair” NH₄‐ion battery of the full‐cell configuration by employing an ammonium Prussian white analogue, (NH₄)_1.47Ni[Fe(CN)₆]_0.88, as the cathode, an organic solid, 3,4,9,10‐perylenetetracarboxylic diimide (PTCDI), as the anode, and 1.0 maqueous (NH₄)₂SO₄as the electrolyte. This novel aqueous ammonium‐ion battery demonstrates encouraging electrochemical performance: an average operation voltage of ca. 1.0 V, an attractive energy density of ca. 43 Wh kg⁻¹based on both electrodes’ active mass, and excellent cycle life over 1000 cycles with 67 % capacity retention. Importantly, the topochemistry results of NH₄⁺in these electrodes point to a new paradigm of NH₄⁺‐based energy storage.

 

The sluggish ion diffusion and electrolyte freezing with volumetric changes limit the low‐temperature performance of rechargeable batteries. Herein, a high‐rate aqueous proton battery (APB) operated at and below −78 °C via a 62 wt% (9.5 m) H₃PO₄electrolyte is reported. The APB is a rocking‐chair battery that operates with protons commuting between a Prussian blue cathode and an MoO₃anode. At −78 °C, the APB full cells exhibit stable cycle life for 450 cycles, high round‐trip efficiency of 85%, and appreciable power performance. The APB delivers 30% of its room‐temperature capacity even at −88 °C. The proton storage mechanism is investigated by ex situ synchrotron XRD, XAS, and XPS. The APB pouch cells demonstrate no capacity fading at −78 °C, and thus offers a safe and reliable candidate for high‐latitude applications.

 

We report reversible electrochemical insertion of NO₃⁻into manganese(II, III) oxide (Mn₃O₄) as a cathode for aqueous dual‐ion batteries. Characterization by TGA, FTIR, EDX, XANES, EXAFS, and EQCM collectively provides unequivocal evidence that reversible oxidative NO₃⁻insertion takes place inside Mn₃O₄. Ex situ HRTEM and corresponding EDX mapping results suggest that NO₃⁻insertion de‐crystallizes the structure of Mn₃O₄. Kinetic studies reveal fast migration of NO₃⁻in the Mn₃O₄structure. This finding may open a new direction for novel low‐cost aqueous dual‐ion batteries.